# Battleships

> Monthly cost of 366 cloud sandbox, VM and agent-runtime providers for a given AI-agent workload, from each provider's public pricing, with features checked against their docs (quoted sources). Built by the boat.dev team; boat.dev is held to the same rules as everyone.

The human page (https://battleships.dev/) is an interactive estimator. Agents: use these static files instead of the HTML. Everything below is plain static files, cached; please don't crawl the HTML page.

## Data
- [Provider index](https://battleships.dev/data/index.json): every provider with category, isolation, features it has, links to its files
- Per provider: `https://battleships.dev/data/providers/<id>.json` (full pricing card, features, quoted evidence) and `https://battleships.dev/data/providers/<id>.md` (readable summary + how it charges)
- [Feature definitions](https://battleships.dev/data/features.json): what each feature means and why it matters
- [Use-case presets](https://battleships.dev/data/presets.json): the workloads behind each ranking
- Rankings per preset, as the page computes them: `https://battleships.dev/data/rankings/<preset>.json` and [all rankings in one page](https://battleships.dev/data/rankings.md)
- [Everything in one text file](https://battleships.dev/llms-full.txt)

## Price your own workload
Download [engine.js](https://battleships.dev/engine.js) and a provider card, then in Node:
```js
require('./engine.js'); const card = require('./e2b.json');
const W = { vcpu: 2, ram: 4, disk: 10, os: 'linux', arch: 'any', gpu: 'none', gpuCount: 1, sessions: 1000, sessionMin: 10, concurrency: 20, alwaysOn: 0,
  cpuUtil: 0.3, ramUtil: 0.5, idleShare: 0.2, cpuPeakUtil: 0.6, ramPeakUtil: 0.7, snapshotGiB: 0, egress: 10, ipv4: 0, seats: 1, persistentDisk: false, pooled: false, classes: null, showZero: false, internet: 'open' };
const r = PM.priceSoft(card, W, { required: ['docker_inside'], maxAccess: 3, region: 'any', idleSuspend: true, idleCapture: 1, overrides: {}, sizing: 'min' });
console.log(r.total, r.breakdown, r.compromises);
```
Fields: sessions a month, minutes per session, machines at once, machines on 24/7, CPU/RAM busy share (0-1), GiB kept, GiB egress.
internet: what sandboxes must reach: 'open' (any site or API), 'pkg' (package registries, git hosting, LLM APIs) or 'none'. Some providers only reach an allowlist on cheap tiers (card.network.internet, plan.internet); the engine then prices the tier that opens the network and reports the prepay or plan it takes (r.access).

## Notes
- Prices are list prices from public pages, re-checked 2026-09-28..30; unknown = not published (never assumed free or unlimited).
- Source code and data: https://github.com/ariana-dot-dev/battleships


# All providers

## Agency Tool Company (agency-tool-company)
https://agencytool.com · dev-env · isolation: bare-metal
- Native product; see regime notes
Features: ≥24 h sessions, full VM (own kernel), Docker inside, GPU, arm64, gVisor or VM (no shared kernel)

## Agent 37 (agent-37)
https://www.agent37.com/pricing · agent-sandbox · isolation: gvisor
- Standard (no dedicated cores): $0.0010958904109589042/vCPU-h, $0.0009589041095890411/GiB-h
- Dedicated Performance: $0.004383561643835617/vCPU-h, $0.0038356164383561643/GiB-h
- Managed OpenClaw/Hermes hosting (flat monthly per instance): Basic $0.005466/h; Plus $0.013685/h; Pro $0.041082/h
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, root, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, custom domains, open internet, SOC 2, Python, Node.js, wake on request, live resize, scheduled runs (cron), spend limits, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, secret proxy

## Agent Relay (agent-relay)
https://agentrelay.com · paas · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, preinstalled agents, self-hosting, open source, macOS, Python, Node.js, webhooks, MCP server, agent harness API, secret proxy

## AgentComputer (agentcomputer)
https://agentcomputer.ai/pricing · agent-sandbox · isolation: firecracker
- On-demand: $0.07/vCPU-h, $0.04375/GiB-h
- macOS QEMU/KVM on-demand: $0.07/vCPU-h, $0.04375/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, desktop GUI, preinstalled agents, SSH, HTTPS preview URLs, open internet, macOS, Python, Node.js, live resize, spend limits, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), shared volumes

## ainclave (ainclave)
https://www.ainclave.com/pricing · agent-sandbox · isolation: firecracker
- On-demand
- Dedicated reserved hosts
Features: idle auto-stop, ≥24 h sessions, full VM (own kernel), egress allowlist, open internet, credential injection, audit logs, EU data residency, spend limits, gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy

## Airtop (airtop)
https://www.airtop.ai/pricing · browser · isolation: vm
- Credit-metered browser session: Credit-metered browser session $null/h
Features: idle auto-stop, ≥24 h sessions, full VM (own kernel), browser, computer-use API, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, SOC 2, HIPAA, Python, Node.js, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel)

## Akamai Cloud / Linode (linode)
https://www.akamai.com/cloud/pricing · hyperscaler · isolation: vm
- Shared CPU: g6-nanode-1 $0.0075/h; g6-standard-1 $0.018/h; g6-standard-2 $0.036/h
- Dedicated CPU: g6-dedicated-2 $0.054/h; g6-dedicated-4 $0.108/h; g6-dedicated-8 $0.216/h
- High-Memory CPU: g7-highmem-1 $0.09/h; g7-highmem-2 $0.18/h; g7-highmem-4 $0.36/h
- GPU RTX 4000 Ada (hourly, no cap): g2-gpu-rtx4000a1-s $0.52/h; g2-gpu-rtx4000a1-m $0.67/h; g2-gpu-rtx4000a1-l $0.96/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, audit logs, EU data residency, scoped API keys, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Alibaba Cloud ACK (Harbor backend) (alibaba-ack)
https://www.alibabacloud.com/product/kubernetes · hyperscaler · isolation: container
- ACK cluster runtime
- ACK pods on virtual nodes (ECI), per vCPU/GiB: $0.02772/vCPU-h, $0.003456/GiB-h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, BYOC, SOC 2, HIPAA, SSO, GPU, arm64, Windows, Python, Node.js, wake on request, EU data residency, MCP server, snapshots on demand, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Alibaba Cloud Agent Sandbox / FC / AgentRun (alibaba-agentrun)
https://help.aliyun.com/zh/agent-sandbox/product-overview/billing-overview · agent-sandbox · isolation: vm
- Economy (per second, allocated): $0.013716/vCPU-h, $0.006876/GiB-h
- Default (per second, allocated): $0.017856/vCPU-h, $0.008928/GiB-h
- Performance (per second, allocated): $0.027324/vCPU-h, $0.013896/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, HTTPS preview URLs, egress allowlist, open internet, SOC 2, HIPAA, SSO, Windows, Python, Node.js, HTTP method/path egress rules, wake on request, live fork (no pause), memory fork, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Alibaba Cloud AgentBay (alibaba-agentbay)
https://www.alibabacloud.com/help/en/agentbay/product-overview/agentbay-billing-instructions · agent-sandbox · isolation: container
- CodeSpace (international), CPU/RAM per second while running: $0.027/vCPU-h, $0.0113/GiB-h
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, anti-bot stealth, CAPTCHA solving, HTTPS preview URLs, egress allowlist, open internet, private networking, SOC 2, HIPAA, SSO, Windows, Python, Node.js, MCP server, hosted agent API (their own agent), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Alibaba Cloud ECS International (alibaba-ecs)
https://www.alibabacloud.com/en/product/ecs/pricing · hyperscaler · isolation: vm
- economy: ecs.e-c1m2.large $0.0281/h; ecs.e-c1m2.xlarge $0.0838/h; ecs.e-c1m4.xlarge $0.1251/h
- compute-c9i: ecs.c9i.large $0.0731/h; ecs.c9i.xlarge $0.1463/h; ecs.c9i.2xlarge $0.2926/h
- burstable-t6: ecs.t6-c1m4.xlarge $0.143/h; ecs.t6-c1m4.2xlarge $0.287/h
- Monthly subscription: ecs.e-c1m2.large $12.96/h; ecs.c9i.large $38.33/h; ecs.e-c1m2.xlarge $38.54/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, live fork (no pause), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Amazon Bedrock AgentCore (Runtime / Code Interpreter / Browser) (aws-agentcore)
https://aws.amazon.com/bedrock/agentcore/pricing/ · hyperscaler · isolation: firecracker
- Code Interpreter session (container; active CPU + peak memory): $0.0895/vCPU-h, $0.00945/GiB-h
- Runtime v1 microVM (same meter as Code Interpreter): $0.0895/vCPU-h, $0.00945/GiB-h
- Runtime v2 consumption (idle memory reclaimed after 120 s): $0.1276/vCPU-h, $0.0169/GiB-h
- Runtime v2 committed baseline (announced, launching by Oct 2026): $0.0997/vCPU-h, $0.0132/GiB-h
Features: persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, computer-use API, code interpreter, browser automation API, desktop control API, browser + desktop control, egress allowlist, open internet, static egress IP, private networking, BYOC, SOC 2, HIPAA, SSO, GPU, arm64, Python, Node.js, wake on request, audit logs, EU data residency, spend limits, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Amazon WorkSpaces Personal (Windows) (aws-workspaces)
https://aws.amazon.com/workspaces/desktop-as-a-service/pricing/ · windows · isolation: vm
- AlwaysOn monthly bundle (licence included, us-east-1): Value (1 vCPU / 2 GB) $0.034247/h; Standard (2 vCPU / 4 GB) $0.045205/h; Performance (2 vCPU / 8 GB) $0.061644/h
- AutoStop: monthly fee + hourly (licence included, us-east-1): Value (1 vCPU / 2 GB) $0.22/h; Standard (2 vCPU / 4 GB) $0.3/h; Performance (2 vCPU / 8 GB) $0.47/h
- AlwaysOn monthly bundle, Ubuntu Linux (us-east-1): Value (1 vCPU / 2 GB) $0.031507/h; Standard (2 vCPU / 4 GB) $0.042466/h; Performance (2 vCPU / 8 GB) $0.058904/h
- AutoStop: monthly fee + hourly, Ubuntu Linux (us-east-1): Value (1 vCPU / 2 GB) $0.19/h; Standard (2 vCPU / 4 GB) $0.28/h; Performance (2 vCPU / 8 GB) $0.45/h
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Windows, wake on request, audit logs, EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Amika (amika)
https://www.amika.dev/pricing · agent-sandbox · isolation: vm
- On-demand: $0.1008/vCPU-h, $0.0324/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, open source, SSO, GPU, Python, Node.js, audit logs, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Amp Orbs (amp)
https://ampcode.com/docs/orbs/sizes-and-costs · dev-env · isolation: vm
- Hosted orb compute, PAYG or subscription allowance: a1.tiny $0.08/h; a1.small $0.17/h; a1.medium $0.33/h
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, HTTPS preview URLs, custom domains, self-hosting, SOC 2, SSO, Python, Node.js, wake on request, scheduled runs (cron), audit logs, spend limits, automatic snapshots, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Anchor Browser (anchor-browser)
https://anchorbrowser.io/pricing · browser · isolation: dedicated-vm
- Cloud browser session (Infrastructure plan: $0.05/browser-hour + $0.01/browser): browser session $0.05/h
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, full VM (own kernel), browser, computer-use API, browser automation API, desktop control API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, egress allowlist, open internet, static egress IP, self-hosting, BYOC, SOC 2, HIPAA, SSO, Python, Node.js, EU data residency, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## Apoxy (apoxy)
https://apoxy.dev · paas · isolation: gvisor
- Native product; see regime notes
Features: ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, HTTPS preview URLs, raw TCP inbound, egress allowlist, open internet, static egress IP, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, arm64, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, scheduled runs (cron), EU data residency, spend limits, MCP server, gVisor or VM (no shared kernel), secret proxy, secret proxy for any API

## Apple Xcode Cloud (xcode-cloud)
https://developer.apple.com/xcode-cloud/ · macos · isolation: apple-vm
- Xcode Cloud compute hours (Xcode build/test only): Xcode Cloud compute hour (Apple-managed, spec not published) $0.4/h
Features: idle auto-stop, ≥24 h sessions, full VM (own kernel), arm64, macOS, Python, webhooks, gVisor or VM (no shared kernel)

## Aptible (aptible)
https://www.aptible.com · paas · isolation: container
- Native product; see regime notes: $0/vCPU-h, $0.08/GiB-h
- CPU-Optimized (C) container profile: $0/vCPU-h, $0.1/GiB-h
- Memory Optimized (R) container profile: $0/vCPU-h, $0.05/GiB-h
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, static egress IP, private networking, self-hosting, BYOC, SOC 2, HIPAA, SSO, Python, Node.js, EU data residency, inbound access rules, secret proxy, extra volumes

## Archal (archal)
https://www.archal.ai/ · agent-sandbox · isolation: container
- Native product; see regime notes: one stateful provider environment $6/h
Features: idle auto-stop, ≥24 h sessions, HTTPS preview URLs, spend limits, MCP server, inbound access rules

## Archil (archil)
https://archil.com/pricing · agent-sandbox · isolation: vm
- Developer plan sandbox, fixed 2 vCPU / 4 GB ($0.27/h after 30 free min): 2vcpu-4gb $0.27/h
- Team plan sandbox, fixed 2 vCPU / 4 GB ($0.18/h after 1,000 included min): 2vcpu-4gb $0.18/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, HTTPS preview URLs, raw TCP inbound, egress allowlist, open internet, BYOC, SOC 2, HIPAA, SSO, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, EU data residency, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Arga Labs (arga-labs)
https://www.argalabs.com/ · agent-sandbox · isolation: container
- Native product; see regime notes
Features: browser, HTTPS preview URLs, self-hosting, BYOC, SSO, Python, Node.js, MCP server, hosted agent API (their own agent), inbound access rules

## Arker (arker)
https://arker.ai/docs/pricing · agent-sandbox · isolation: vm
- AWS US VMs, on-demand run time (docs planning-table rates): small (2 vCPU / 8 GiB / 32 GiB) $0.187736/h; large (4 vCPU / 16 GiB / 32 GiB) $0.372016/h
- AWS US pool, 12-month commitment (always-on, 19.55% off): small (2 vCPU / 8 GiB / 32 GiB) $0.15104/h; large (4 vCPU / 16 GiB / 32 GiB) $0.299299/h
- AWS US pool, 24-month commitment (always-on, 22.83% off): small (2 vCPU / 8 GiB / 32 GiB) $0.144871/h; large (4 vCPU / 16 GiB / 32 GiB) $0.287075/h
- AWS US pool, 36-month commitment (always-on, 48.56% off): small (2 vCPU / 8 GiB / 32 GiB) $0.096581/h; large (4 vCPU / 16 GiB / 32 GiB) $0.191384/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, systemd, browser, desktop GUI, computer-use API, code interpreter, browser automation API, desktop control API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, private networking, BYOC, SSO, GPU, arm64, Python, Node.js, Go, Rust, GPU desktop, credential injection, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, EU data residency, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Artillery (artillery)
https://www.artillery.io/pricing · dev-env · isolation: container
- Test workers and Cloud platform
Features: ≥24 h sessions, browser, public IPv4, egress allowlist, self-hosting, BYOC, open source, SSO, Node.js, audit logs, EU data residency

## AutoComputer (autocomputer)
https://www.autocomputer.ai/ · windows · isolation: firecracker
- Native product; see regime notes
Features: ≥24 h sessions, full VM (own kernel), browser, desktop GUI, computer-use API, browser + desktop control, HTTPS preview URLs, Windows, Python, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules

## AWS CodeBuild (aws-codebuild)
https://aws.amazon.com/codebuild/pricing/ · paas · isolation: container
- On-demand EC2 Linux general1.small: general1.small (BUILD_GENERAL1_SMALL) $0.3/h
- On-demand EC2 Linux general1 (medium..2xlarge): general1.medium $0.6/h; general1.large $1.2/h; general1.xlarge $4.788/h
- On-demand EC2 Arm (arm1): arm1.small $0.204/h; arm1.medium $0.42/h; arm1.large $0.9/h
- Interactive CodeBuild Sandbox Linux GPU: gpu1.small $1.152/h; gpu1.large $6.516/h
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, root, SSH, egress allowlist, open internet, static egress IP, private networking, self-hosting, SOC 2, HIPAA, SSO, GPU, arm64, Windows, macOS, Python, Node.js, EU data residency, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## AWS EC2 (reference VMs) (aws-ec2)
https://aws.amazon.com/ec2/pricing/on-demand/ · hyperscaler · isolation: vm
- x86 on-demand c7i/m7i (us-east-1): c7i.large $0.08925/h; m7i.large $0.1008/h; c7i.xlarge $0.1785/h
- x86 on-demand c7i/m7i (eu-north-1, cheapest EU checked): c7i.large $0.09555/h; m7i.large $0.1071/h; c7i.xlarge $0.1911/h
- x86 on-demand c7i/m7i (ap-south-1, cheapest Asia checked): c7i.large $0.08925/h; m7i.large $0.10605/h; c7i.xlarge $0.1785/h
- Graviton3 on-demand c7g/m7g (us-east-1): c7g.large $0.0725/h; m7g.large $0.0816/h; c7g.xlarge $0.145/h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, Python, HTTP method/path egress rules, EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## AWS EC2 (Windows Server) (aws-ec2-windows)
https://aws.amazon.com/ec2/pricing/on-demand/ · windows · isolation: vm
- On-Demand c7i/m7i + Windows Server licence (us-east-1): c7i.large $0.08925/h; m7i.large $0.1008/h; c7i.xlarge $0.1785/h
- t3 burstable, Windows licence included (us-east-1): t3.medium $0.06/h; t3.large $0.1108/h; t3.xlarge $0.24/h
- c7i-flex/m7i-flex Windows (us-east-1): c7i-flex.large $0.17219/h; m7i-flex.large $0.18316/h; c7i-flex.xlarge $0.34438/h
- spot
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, macOS, HTTP method/path egress rules, live fork (no pause), audit logs, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## AWS EC2 Mac (Dedicated Host) (aws-ec2-mac)
https://aws.amazon.com/ec2/instance-types/mac/ · macos · isolation: dedicated-host
- On-Demand Dedicated Host, whole Mac (us-east-1): mac2.metal (Apple M1, 8 vCPU / 16 GiB) $0.65/h; mac2-m2.metal (Apple M2, 8 vCPU / 24 GiB) $0.878/h; mac-m4.metal (Apple M4, 10 vCPU / 24 GiB) $1.23/h
- Host split into 2 macOS VMs (Apple SLA cap), per-VM price: 1/2 of mac2.metal (Apple M1, 8 vCPU / 16 GiB) $0.325/h; 1/2 of mac2-m2.metal (Apple M2, 8 vCPU / 24 GiB) $0.439/h; 1/2 of mac-m4.metal (Apple M4, 10 vCPU / 24 GiB) $0.615/h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), nested virtualization, root, browser, desktop GUI, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, arm64, macOS, Python, HTTP method/path egress rules, audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## AWS Fargate (aws-fargate)
https://aws.amazon.com/fargate/pricing/ · hyperscaler · isolation: firecracker
- Linux/x86 on-demand (us-east-1 = eu-west-1): $0.04048/vCPU-h, $0.004445/GiB-h
- Linux/x86 on-demand (ap-south-1 Mumbai, cheapest Asia): $0.04256/vCPU-h, $0.004655/GiB-h
- Linux/ARM64 Graviton on-demand (us-east-1 = eu-west-1): $0.03238/vCPU-h, $0.00356/GiB-h
- Linux/ARM64 on-demand (ap-south-1, cheapest region overall): $0.02383/vCPU-h, $0.00261/GiB-h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, public IPv4, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, static egress IP, private networking, SOC 2, HIPAA, SSO, arm64, Windows, HTTP method/path egress rules, webhooks, audit logs, EU data residency, spend limits, MCP server, gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## AWS Lambda (aws-lambda)
https://aws.amazon.com/lambda/pricing/ · hyperscaler · isolation: firecracker
- x86 functions, duration tier 1: $0/vCPU-h, $0.06/GiB-h
- Arm64 (Graviton) functions, duration tier 1: $0/vCPU-h, $0.048/GiB-h
- Compute Savings Plan 1y/3y no-upfront, x86: $0/vCPU-h, $0.05292/GiB-h
- Compute Savings Plan no-upfront, Arm64: $0/vCPU-h, $0.04212/GiB-h
Features: volumes, idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), HTTPS preview URLs, egress allowlist, open internet, private networking, SOC 2, HIPAA, SSO, arm64, Python, Node.js, Java, wake on request, scheduled runs (cron), audit logs, EU data residency, MCP server, gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## AWS Lambda MicroVMs (aws-lambda-microvms)
https://aws.amazon.com/lambda/pricing/ · hyperscaler · isolation: firecracker
- Arm64 (Graviton) baseline, us-east-1/us-east-2/us-west-2/ap-south-1: $0.0997/vCPU-h, $0.0132/GiB-h
- Arm64 baseline, eu-north-1 (cheapest EU): $0.09975/vCPU-h, $0.013208/GiB-h
- x86 baseline, us-east-1/us-east-2/us-west-2/ap-south-1: $0.11756/vCPU-h, $0.015565/GiB-h
- x86 baseline, eu-north-1: $0.117954/vCPU-h, $0.015617/GiB-h
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, arm64, HTTP method/path egress rules, wake on request, EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Azure Container Apps (Consumption, Dedicated, Dynamic Sessions) (azure-container-apps)
https://azure.microsoft.com/en-us/pricing/details/container-apps/ · hyperscaler · isolation: vm
- Consumption plan, active rate (East US / North Europe / Japan East class): $0.0864/vCPU-h, $0.0108/GiB-h
- Serverless GPU workload profile (Consumption, one GPU per replica): $0.0864/vCPU-h, $0.0108/GiB-h
- Consumption + 1y savings plan: $0.07344/vCPU-h, $0.00918/GiB-h
- Consumption + 3y savings plan: $0.071712/vCPU-h, $0.008964/GiB-h
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, root, code interpreter, HTTPS preview URLs, custom domains, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, audit logs, EU data residency, MCP server, inbound access rules, extra volumes, shared volumes

## Azure Container Apps Sandboxes (azure-container-apps-sandboxes)
https://learn.microsoft.com/en-us/azure/container-apps/sandboxes-overview · hyperscaler · isolation: vm
- Sandbox tiers XS-XL + programmatic sizes up to 16 vCPU / 32 GB (Consumption meters): XS $0.027/h; S $0.054/h; M $0.108/h
- Sandbox tiers + 1y savings plan (assumed to apply): XS $0.02295/h; S $0.0459/h; M $0.0918/h
- Sandbox tiers + 3y savings plan (assumed to apply): XS $0.02241/h; S $0.04482/h; M $0.08964/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, code interpreter, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, private networking, SOC 2, HIPAA, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live fork (no pause), memory fork, webhooks, audit logs, EU data residency, spend limits, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Azure Pipelines (azure-pipelines)
https://azure.microsoft.com/en-us/pricing/details/devops/azure-devops-services/ · paas · isolation: vm
- Microsoft-hosted parallel jobs ($40 each/month, unlimited minutes)
- Microsoft-hosted macOS parallel jobs ($40 each/month, unlimited minutes)
- GitHub-hosted agents (pay-as-you-go per minute, macOS arm64, preview): macOS Standard (M1) $3.72/h; macOS XL (M2) $6.12/h
- Managed DevOps Pools (your Azure VMs + $15 self-hosted parallel job)
Features: persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, egress allowlist, open internet, self-hosting, SOC 2, HIPAA, SSO, arm64, Windows, macOS, Python, Node.js, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Azure Virtual Machines (Linux) (azure-vm)
https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/ · hyperscaler · isolation: vm
- Pay-as-you-go Linux Dsv5/Dasv5/Dlsv5/Fsv2 (East US): D2s_v5 $0.096/h; D4s_v5 $0.192/h; D8s_v5 $0.384/h
- Pay-as-you-go Linux Dpsv5 Arm (Ampere Altra) (East US): D2ps_v5 $0.077/h; D4ps_v5 $0.154/h; D8ps_v5 $0.308/h
- B-series v2 burstable Basv2 'l' sizes, 30% baseline (East US): B2als_v2 $0.0376/h; B4als_v2 $0.133/h; B8als_v2 $0.266/h
- B-series v2 burstable Basv2, 40% baseline (East US): B2as_v2 $0.0752/h; B4as_v2 $0.15/h; B8as_v2 $0.301/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, HTTP method/path egress rules, webhooks, audit logs, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Azure Virtual Machines (Windows Server) (azure-windows)
https://azure.microsoft.com/en-us/pricing/details/virtual-machines/windows/ · windows · isolation: vm
- Pay-as-you-go D/F-series + Windows licence (East US): Standard_F2s_v2 $0.085/h; Standard_D2as_v5 $0.086/h; Standard_D2s_v5 $0.096/h
- B-series burstable + Windows licence (East US): Standard_B2s $0.0416/h; Standard_B2ms $0.0832/h; Standard_B4ms $0.166/h
- Azure Spot VM (representative East US): Standard_D4als_v7 $0.06693/h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, HTTP method/path egress rules, webhooks, audit logs, EU data residency, spend limits, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Baponi (baponi)
https://baponi.ai/pricing/ · agent-sandbox · isolation: container
- On-demand: $0.06/vCPU-h, $0/GiB-h
Features: snapshots, persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, code interpreter, preinstalled agents, egress allowlist, open internet, self-hosting, BYOC, SSO, Python, Node.js, wake on request, audit logs, EU data residency, scoped API keys, MCP server, automatic snapshots, extra volumes, shared volumes

## Baseten (baseten)
https://docs.baseten.co/deployment/resources · gpu-cloud · isolation: container
- T4, serverless (×1): $0/vCPU-h, $0/GiB-h
- L4, serverless (×1): $0/vCPU-h, $0/GiB-h
- A10G, serverless (×1): $0/vCPU-h, $0/GiB-h
- A100-80GB, serverless (×1): $0/vCPU-h, $0/GiB-h
Features: volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, SSH, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, SOC 2, HIPAA, SSO, GPU, Python, wake on request, EU data residency, MCP server, inbound access rules, extra volumes, shared volumes

## Beam (beam)
https://www.beam.cloud/pricing · agent-sandbox · isolation: gvisor
- Sandbox (CPU), per-ms on allocated cores + RAM: $0.0675/vCPU-h, $0.02304/GiB-h
- Sandbox with GPU (rate not published): $0.189/vCPU-h, $0.0198/GiB-h
- Serverless function / pod (not a sandbox), 1/3 of sandbox CPU rate: $0.0225/vCPU-h, $0.00756/GiB-h
- On-demand GPU machine, 1x RTX 4090 (8 vCPU / 64 GiB bundled): $0/vCPU-h, $0/GiB-h
Features: snapshots, memory snapshots, fork/clone, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, Docker inside, root, SSH, HTTPS preview URLs, raw TCP inbound, egress allowlist, open internet, static egress IP, self-hosting, BYOC, open source, SOC 2, HIPAA, GPU, Python, Node.js, live fork (no pause), memory fork, webhooks, scheduled runs (cron), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## Bitbucket Pipelines (bitbucket-pipelines)
https://www.atlassian.com/software/bitbucket/pricing · paas · isolation: container
- Atlassian cloud runners 1x/2x: size 1x $0.6/h; size 2x $1.2/h
- Atlassian cloud runners 4x..32x (paid plans): size 4x $2.4/h; size 8x $4.8/h; size 16x $9.6/h
Features: idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, root, preinstalled agents, open internet, self-hosting, SOC 2, SSO, arm64, Python, Node.js, audit logs, spend limits, MCP server, agent harness API

## Bitrise (mobile CI) (bitrise)
https://bitrise.io/pricing · macos · isolation: vm
- Starter plan: flat $99/$199/$299 per month (1/2/4 concurrent Medium builds), unlimited usage
- Per-minute 'Zero Margin' machine cost, macOS: macOS Medium (M4 5 CPU / 6 GB) $0.432/h; macOS Large (M4 5 CPU / 14 GB) $0.576/h; macOS 4Large (M4 Pro 7 CPU / 27 GB) $0.864/h
- Per-minute machine cost, Linux: Linux 2S $0.132/h; Linux M $0.264/h; Linux 2M $0.396/h
- Remote Dev Environments, macOS VM (per-minute, monthly cap per VM): RDE Basic macOS M2 Pro Large $1.728/h; RDE Pro macOS M2 Pro X Large $3.456/h
Features: pause/resume, persistent disk, idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, static egress IP, self-hosting, BYOC, SOC 2, SSO, arm64, macOS, Python, Node.js, webhooks, EU data residency, MCP server, agent harness API, gVisor or VM (no shared kernel), firewall inside (nftables)

## Blacksmith (GitHub Actions runners) (blacksmith)
https://www.blacksmith.sh/pricing · macos · isolation: vm
- macOS M4 runners, per minute: blacksmith-6vcpu-macos (M4, 6 vCPU / 24 GB) $4.8/h; blacksmith-12vcpu-macos (M4, 12 vCPU / 48 GB) $9.6/h
- Windows Server 2025 runners (public beta), per minute: 2 vCPU / 7 GB $0.48/h; 4 vCPU / 14 GB $0.96/h; 8 vCPU / 28 GB $1.92/h
- Ubuntu x64 runners, per minute: 2 vCPU / 8 GB $0.24/h; 4 vCPU / 16 GB $0.48/h; 8 vCPU / 32 GB $0.96/h
- Ubuntu ARM runners, per minute: blacksmith-2vcpu-ubuntu-2404-arm (2 vCPU / 6 GB) $0.15/h; blacksmith-4vcpu-ubuntu-2404-arm (4 vCPU / 12 GB) $0.3/h; blacksmith-8vcpu-ubuntu-2404-arm (8 vCPU / 24 GB) $0.6/h
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, SSH, static egress IP, SOC 2, arm64, Windows, macOS, Python, Node.js, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes

## Blaxel (blaxel)
https://blaxel.ai/pricing · agent-sandbox · isolation: firecracker
- Sandbox (active time, RAM-only pricing, CPU bundled): 2 GB $0.0828/h; 4 GB $0.1656/h; 8 GB $0.3312/h
- Batch Jobs (non-interactive tasks, max 24 h): $0/vCPU-h, $0.0216/GiB-h
- Agent Runtime (announced; unpriced)
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, preinstalled agents, HTTPS preview URLs, custom domains, egress allowlist, open internet, static egress IP, private networking, BYOC, SOC 2, HIPAA, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live fork (no pause), memory fork, scheduled runs (cron), EU data residency, zero data retention, scoped API keys, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## boat.dev (boat)
https://docs.boat.dev/pricing · agent-sandbox · isolation: vm
- On-demand (flat per size, per second while running): small $0.018/h; default $0.036/h; large $0.072/h
Features: snapshots, fork/clone, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, anti-bot stealth, preinstalled agents, SSH, HTTPS preview URLs, raw TCP inbound, open internet, Python, Node.js, Go, Rust, Java, live fork (no pause), webhooks, EU data residency, zero data retention, scoped API keys, spend limits, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## boxd (boxd)
https://boxd.sh/pricing · agent-sandbox · isolation: vm
- Pay as you go (credits, USD list): $0.059/vCPU-h, $0.018/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, browser, desktop GUI, anti-bot stealth, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, self-hosting, BYOC, SSO, Python, Node.js, Go, credential injection, wake on request, live fork (no pause), memory fork, scheduled runs (cron), audit logs, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes

## boxes.dev (boxes-dev)
https://boxes.dev/ · dev-env · isolation: firecracker
- Starter marginal box-hours: Starter-default $0.75/h; Starter-4x16 $1.0425/h; Starter-8x16 $1.5/h
- Pro marginal box-hours: Pro-default $0.6/h; Pro-4x16 $0.834/h; Pro-8x16 $1.2/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, preinstalled agents, SSH, HTTPS preview URLs, open internet, BYOC, SSO, wake on request, live fork (no pause), scheduled runs (cron), audit logs, spend limits, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Bright Data Browser API (bright-data-browser)
https://brightdata.com/pricing/scraping-browser · browser · isolation: not published
- Bandwidth-billed browser: Bandwidth-billed browser $0/h
Features: idle auto-stop, browser, computer-use API, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, SOC 2, SSO, Python, Node.js, MCP server

## Brimble Sandboxes (brimble)
https://brimble.io/pricing · agent-sandbox · isolation: gvisor
- On-demand
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, custom image (Docker or snapshot), start from your own snapshot, Docker inside, code interpreter, preinstalled agents, egress allowlist, Python, Node.js, live fork (no pause), EU data residency, spend limits, MCP server, snapshots on demand, gVisor or VM (no shared kernel), extra volumes

## Browser Use Cloud (browser-use)
https://browser-use.com/pricing · browser · isolation: not published
- Browser runtime: Browser runtime $0.02/h
Features: persistent disk, idle auto-stop, browser, computer-use API, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, egress allowlist, open internet, self-hosting, open source, SOC 2, HIPAA, Python, Node.js, EU data residency, MCP server, hosted agent API (their own agent)

## BrowserAct (browseract)
https://www.browseract.com/pricing · browser · isolation: not published
- Cloud Bot managed browser: Cloud Bot managed browser $null/h
Features: pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, browser, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, static egress IP, Python, MCP server, hosted agent API (their own agent)

## Browserbase (browserbase)
https://www.browserbase.com/pricing · browser · isolation: vm
- Browser session, Developer overage rate ($0.12/browser-hour): browser-session $0.12/h
- Browser session, Startup overage rate ($0.10/browser-hour): browser-session $0.1/h
- Scale plan (custom usage-based rates, Verified/advanced stealth): browser-session $null/h
Features: persistent disk, idle auto-stop, full VM (own kernel), browser, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, egress allowlist, open internet, static egress IP, SOC 2, HIPAA, SSO, Python, Node.js, webhooks, EU data residency, zero data retention, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel), secret proxy

## BrowserCloud (browsercloud)
https://browsercloud.io/pricing · browser · isolation: container
- Hobby plan browser-hours (120 h for $19): Hobby browser session $0.1583/h
- Startup plan browser-hours (1000 h for $99): Startup browser session $0.099/h
- Business plan browser-hours (3000 h for $249): Business browser session $0.083/h
- Unlimited plan: dedicated cluster, flat monthly per parallel sessions
Features: persistent disk, ≥24 h sessions, browser, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, static egress IP, SOC 2, Python, Node.js

## Browserless (browserless)
https://www.browserless.io/pricing · browser · isolation: container
- Free quota only: Free quota only $null/h
- Prototyping unit overage: Prototyping $0.24/h
- Starter unit overage: Starter $0.204/h
- Scale unit overage: Scale $0.18/h
Features: persistent disk, idle auto-stop, browser, computer-use API, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, egress allowlist, open internet, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, Python, Node.js, HTTP method/path egress rules, webhooks, EU data residency, MCP server, snapshots on demand, hosted agent API (their own agent), inbound access rules

## Buddy Sandboxes (buddy)
https://buddy.works/pricing · agent-sandbox · isolation: vm
- On-demand: $0.0144/vCPU-h, $0.0144/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, SOC 2, SSO, Python, Node.js, Go, wake on request, live resize, live fork (no pause), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## BuildJet (buildjet)
https://buildjet.com/for-github-actions · dev-env · isolation: not published · discontinued

Features: ≥24 h sessions

## Buildkite hosted agents (buildkite-hosted)
https://buildkite.com/pricing · macos · isolation: vm
- macOS M4 hosted agents, $0.02 per vCPU-minute: M4 Medium (6 vCPU / 28 GB) $7.2/h; M4 Large (12 vCPU / 56 GB) $14.4/h
- Hosted Linux agents (Small..Large): Small LINUX_AMD64_2X4 $0.48/h; Medium LINUX_AMD64_4X16 $0.96/h; Large LINUX_AMD64_8X32 $1.92/h
- Hosted Linux agents XL..3XL (Enterprise): XL LINUX_AMD64_16X64 $3.84/h; 2XL LINUX_AMD64_32X128 $7.68/h; 3XL LINUX_AMD64_64X256 $15.36/h
- Windows hosted agents (private preview): Small WINDOWS_AMD64_2X8 $null/h; Medium WINDOWS_AMD64_4X16 $null/h; Large WINDOWS_AMD64_8X32 $null/h
Features: snapshots, fork/clone, persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, browser, desktop GUI, SSH, static egress IP, self-hosting, SOC 2, SSO, arm64, Windows, macOS, Python, Node.js, webhooks, MCP server, automatic snapshots, gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy, extra volumes

## Butter (butter)
https://butter.dev · agent-sandbox · isolation: gvisor
- Native product; see regime notes
Features: ≥24 h sessions, self-hosting, open source, arm64, Python, Node.js, gVisor or VM (no shared kernel)

## Bytebot (bytebot)
https://www.bytebot.ai/ · agent-sandbox · isolation: container
- Self-hosted open-source Docker desktop agent
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, root, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, open internet, self-hosting, open source, arm64, Python, Node.js, EU data residency, MCP server, hosted agent API (their own agent), firewall inside (nftables), extra volumes

## Capy (capy)
https://capy.ai · agent-sandbox · isolation: vm
- Per-thread VM, billed hourly by size while awake: Small $0.1/h; Medium $0.2/h; Large $0.4/h
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, browser, desktop GUI, HTTPS preview URLs, egress allowlist, open internet, private networking, SOC 2, SSO, Python, Node.js, spend limits, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables)

## Caution (caution)
https://caution.co/pricing.html · paas · isolation: vm
- Fully managed enclave
- Managed enclave in own AWS
- AGPL self-host
Features: ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), SSH, public IPv4, HTTPS preview URLs, custom domains, raw TCP inbound, static egress IP, self-hosting, BYOC, open source, EU data residency, gVisor or VM (no shared kernel), inbound access rules

## Cedana (cedana)
https://cedana.com/ · other (checkpoint/restore software; not a compute provider) · isolation: not published
- Native product; see regime notes
Features: snapshots, memory snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, self-hosting, open source, arm64, automatic snapshots, snapshots on demand, extra volumes

## Celesto Cloud (celesto)
https://celesto.ai/pricing · agent-sandbox · isolation: vm
- Current PAYG fixed sizes(shared CPU): Micro $0.03/h; Small $0.04/h; Standard $0.06/h
Features: snapshots, fork/clone, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), browser, SSH, HTTPS preview URLs, egress allowlist, open internet, self-hosting, open source, Python, Node.js, live fork (no pause), automatic snapshots, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Cerebrium (cerebrium)
https://cerebrium.ai/pricing · gpu-cloud · isolation: container
- Interruptible: $0.02358/vCPU-h, $0.007992/GiB-h
- Protected compute: $0.02358/vCPU-h, $0.007992/GiB-h
- B200, serverless: $0.02358/vCPU-h, $0.007992/GiB-h
- H200, serverless: $0.02358/vCPU-h, $0.007992/GiB-h
Features: snapshots, memory snapshots, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, HTTPS preview URLs, custom domains, private networking, SOC 2, HIPAA, GPU, Python, wake on request, webhooks, audit logs, EU data residency, snapshots on demand, inbound access rules, extra volumes, shared volumes

## Chronicle Labs (chronicle-labs)
https://chronicle-labs.com · agent-sandbox · isolation: qemu-kvm
- Native service
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, egress allowlist, arm64, Python, Node.js, live fork (no pause), memory fork, scoped API keys, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes

## CircleCI (circleci)
https://circleci.com/pricing/price-list/ · macos · isolation: vm
- macOS M4 Pro, credits per minute: m4pro.medium (6 CPU / 28 GB, 200 cr/min) $7.2/h; m4pro.large (12 CPU / 56 GB, 400 cr/min) $14.4/h
- Windows VM, credits per minute: windows.medium (4 CPU / 16 GB, 40 cr/min) $1.44/h; windows.large (8 CPU / 32 GB, 120 cr/min) $4.32/h; windows.xlarge (16 CPU / 64 GB, 210 cr/min) $7.56/h
- Linux VM (machine) executor (x86), credits per minute: machine medium (2 CPU / 7.5 GB, 10 cr/min) $0.36/h; machine large (4 CPU / 15 GB, 20 cr/min) $0.72/h; machine xlarge (8 CPU / 32 GB, 100 cr/min) $3.6/h
- Linux Docker executor (x86), credits per minute: docker small (1 CPU / 2 GB, 5 cr/min) $0.18/h; docker medium (2 CPU / 4 GB, 10 cr/min) $0.36/h; docker medium+ (3 CPU / 6 GB, 15 cr/min) $0.54/h
Features: snapshots, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, browser, SSH, open internet, static egress IP, self-hosting, SOC 2, SSO, GPU, arm64, Windows, macOS, Python, Node.js, webhooks, audit logs, MCP server, snapshots on demand, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables)

## Cirrus Runners (cirrus-runners)
https://cirrus-runners.app/pricing/ · macos · isolation: apple-vm · discontinued

Features: idle auto-stop, ≥24 h sessions, nested virtualization, browser, SOC 2, GPU, arm64, macOS, firewall inside (nftables)

## Civo Compute (civo)
https://www.civo.com/pricing · hyperscaler · isolation: vm
- Standard Compute: xSmall $0.00744/h; Small $0.014881/h; Medium $0.029762/h
- Performance Compute: Performance-S $0.119055/h; Performance-M $0.238096/h; Performance-L $0.476192/h
- CPU optimized: CPU-8 $0.190476/h; CPU-16 $0.380952/h; CPU-32 $0.761905/h
- RAM Optimized Compute: RAM-S $0.107137/h; RAM-M $0.214288/h; RAM-L $0.428575/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, GPU, Python, Node.js, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Claude Managed Agents (claude-managed-agents)
https://platform.claude.com/docs/en/about-claude/pricing#claude-managed-agents-pricing · agent-sandbox · isolation: container
- Cloud sandbox session runtime ($0.08 per running session-hour, flat): cloud sandbox (fixed) $0.08/h
- Self-hosted sandbox environment (your infra, Anthropic orchestration): self-hosted worker $null/h
- Messages API code execution tool container (not Managed Agents): code execution container $0.05/h
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, browser, preinstalled agents, egress allowlist, open internet, self-hosting, BYOC, SOC 2, SSO, Python, Node.js, Go, Rust, Java, credential injection, wake on request, webhooks, scheduled runs (cron), spend limits, MCP server, automatic snapshots, agent harness API, secret proxy, secret proxy for any API, extra volumes, shared volumes

## Claw 2 Agent (claw2agent)
https://claw2agent.com/pricing · agent-sandbox · isolation: dedicated-vm
- Managed workspace: Managed workspace $null/h
Features: persistent disk, ≥24 h sessions, full VM (own kernel), browser, code interpreter, preinstalled agents, BYOC, Python, Node.js, credential injection, spend limits, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy

## Clever Cloud (clever-cloud)
https://www.clever.cloud/pricing/ · paas · isolation: vm
- Docker application, Paris: nano $0.00725/h; XS $0.019333333333/h; S $0.038666666667/h
Features: pause/resume, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, SSH, HTTPS preview URLs, raw TCP inbound, static egress IP, self-hosting, Python, Node.js, EU data residency, MCP server, gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## CloudAxis (cloudaxis)
https://cloudaxis.ai/pricing/ · browser · isolation: not published
- Capped browser quota: Capped browser quota $null/h
Features: persistent disk, ≥24 h sessions, browser, computer-use API, code interpreter, browser automation API, browser + desktop control, residential IPs, Python, Node.js, scheduled runs (cron), spend limits, hosted agent API (their own agent)

## CloudBrowser AI (cloudbrowser)
https://cloudbrowser.ai/ · browser · isolation: container
- Browser-hour overage unpublished: Browser-hour overage unpublished $null/h
Features: idle auto-stop, ≥24 h sessions, browser, desktop GUI, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, open internet, static egress IP, Python, Node.js, EU data residency, MCP server

## CloudCruise (cloudcruise)
https://cloudcruise.com/pricing · browser · isolation: dedicated-vm
- Automation hours ($3/h after 2 free hours, Starter): automation hour (VM, proxies and AI included) $3/h
Features: ≥24 h sessions, full VM (own kernel), browser, browser automation API, anti-bot stealth, CAPTCHA solving, egress allowlist, open internet, SOC 2, HIPAA, SSO, Python, Node.js, webhooks, audit logs, hosted agent API (their own agent), gVisor or VM (no shared kernel)

## Cloudflare Browser Run / Kitesurf (cloudflare-browser)
https://developers.cloudflare.com/browser-run/pricing/ · browser · isolation: container
- Chromium Browser Sessions: Chromium Browser Sessions $0.09/h
- Quick Actions: Quick Actions $0.09/h
- Kitesurf beta: Kitesurf beta $0/h
Features: idle auto-stop, ≥24 h sessions, browser, computer-use API, browser automation API, egress allowlist, open internet, SOC 2, HIPAA, SSO, Python, Node.js, MCP server

## Cloudflare Sandbox SDK / Containers (cloudflare-sandbox)
https://developers.cloudflare.com/containers/platform/pricing/ · agent-sandbox · isolation: vm
- Pay-as-you-go (active CPU, provisioned RAM/disk): $0.072/vCPU-h, $0.009/GiB-h
Features: snapshots, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, code interpreter, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, open source, SOC 2, HIPAA, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, audit logs, EU data residency, snapshots on demand, agent harness API, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, secret proxy, secret proxy for any API, extra volumes, shared volumes

## Cloudflare Workers for Platforms (cloudflare-workers-for-platforms)
https://developers.cloudflare.com/cloudflare-for-platforms/workers-for-platforms/reference/pricing/ · paas · isolation: isolate
- User Workers (V8 isolates), per request + CPU-ms: $0.072/vCPU-h, $0/GiB-h
- Dynamic Workers (Worker Loader isolates), Workers Paid plan: $0.072/vCPU-h, $0/GiB-h
Features: idle auto-stop, ≥24 h sessions, HTTPS preview URLs, custom domains, egress allowlist, open internet, open source, SOC 2, HIPAA, SSO, Python, Node.js, HTTP method/path egress rules, wake on request, scheduled runs (cron), audit logs, EU data residency, scoped API keys, MCP server, gVisor or VM (no shared kernel), inbound access rules, secret proxy, secret proxy for any API

## cloudrouter (Manaflow) (cloudrouter)
https://cloudrouter.dev · agent-sandbox · isolation: firecracker · discontinued

Features: snapshots, memory snapshots, pause/resume, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, SSH, HTTPS preview URLs, open source, GPU, Python, Node.js, gVisor or VM (no shared kernel), firewall inside (nftables)

## Clusy (clusy)
https://www.clusy.io/pricing · dev-env · isolation: firecracker
- On-demand: 
Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), browser, code interpreter, SSO, GPU, Python, wake on request, live fork (no pause), memory fork, automatic snapshots, gVisor or VM (no shared kernel)

## Coasty (coasty)
https://coasty.ai/pricing · agent-sandbox · isolation: vm
- Linux running default: Linux running default $0.05/h
- Windows running default: Windows running default $0.09/h
Features: snapshots, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, residential IPs, SSH, public IPv4, open internet, BYOC, SOC 2, HIPAA, SSO, Windows, macOS, webhooks, scoped API keys, spend limits, MCP server, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables)

## Codemagic (mobile CI) (codemagic)
https://codemagic.io/pricing/ · macos · isolation: vm
- Pay-as-you-go macOS minutes: Mac mini M2 (8-core CPU / 12 GB) $5.7/h; Mac mini M4 (10-core CPU / 16 GB) $6.84/h
- Pay-as-you-go Windows minutes: Windows VM (8 vCPU / 32 GB) $2.7/h
- Fixed price annual plan: unlimited minutes, 3 concurrent builds
- Pay-as-you-go Linux X2 minutes: Linux X2 (8 vCPU / 32 GB) $2.7/h
Features: idle auto-stop, full VM (own kernel), Docker inside, nested virtualization, browser, desktop GUI, SSH, static egress IP, SOC 2, SSO, arm64, Windows, macOS, Python, Node.js, webhooks, audit logs, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Coder (coder)
https://coder.com/pricing · dev-env · isolation: not published
- Self-hosted workspaces
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, desktop GUI, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, open source, SOC 2, SSO, GPU, arm64, Windows, macOS, Python, Node.js, HTTP method/path egress rules, audit logs, EU data residency, spend limits, MCP server, agent harness API, inbound access rules, firewall inside (nftables), secret proxy, extra volumes

## CodeSandbox SDK (codesandbox-sdk)
https://codesandbox.io/docs/sdk/pricing · agent-sandbox · isolation: firecracker
- VM credits, fixed sizes ($0.01486/credit, per started minute): Pico $0.0743/h; Nano $0.1486/h; Micro $0.2972/h
- Enterprise bulk VM credit packs (up to 50% off): Pico $0.03715/h; Nano $0.0743/h; Micro $0.1486/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, code interpreter, HTTPS preview URLs, custom domains, SOC 2, SSO, Python, Node.js, Go, Rust, wake on request, live resize, live fork (no pause), memory fork, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Collimate (collimate)
https://collimate.ai/pricing · agent-sandbox · isolation: firecracker
- On-demand: $0.04/vCPU-h, $0.012/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), browser, code interpreter, egress allowlist, open internet, self-hosting, BYOC, SSO, Python, Node.js, live fork (no pause), memory fork, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules

## Comfy Deploy (comfy-deploy)
https://app.comfydeploy.com/pricing · gpu-cloud · isolation: not published · discontinued

Features: volumes, idle auto-stop, ≥24 h sessions, self-hosting, open source, GPU, wake on request, extra volumes, shared volumes

## Computer Use Cloud (computeruse.run) (computeruse-cloud)
https://computeruse.run/ · agent-sandbox · isolation: not published · discontinued

Features: ≥24 h sessions

## ComputerUse.space (computeruse-space)
https://computeruse.space/ · agent-sandbox · isolation: dedicated-host
- ThinkCentre 8 GB planned: ThinkCentre 8 GB planned $null/h
- ThinkCentre 16 GB planned: ThinkCentre 16 GB planned $null/h
- ThinkCentre annual prepay (planned): Standard 8 GB annual $null/h; Creator 16 GB annual $null/h
Features: persistent disk, ≥24 h sessions, full VM (own kernel), browser, desktop GUI, computer-use API, desktop control API, browser + desktop control, residential IPs, MCP server, gVisor or VM (no shared kernel), firewall inside (nftables)

## Contabo (contabo)
https://contabo.com/en-us/pricing/ · hyperscaler · isolation: vm
- Core monthly list: Cloud VPS 18 $58.8/h; Cloud VPS 16 $44.5/h; Cloud VPS 12 $30/h
- Performance monthly list: Cloud VPS Plus 4 $16.25/h; Cloud VPS Plus 18 $119/h; Cloud VPS Plus 16 $95/h
- Max Performance VPS (VDS) monthly list: Cloud VDS S $59/h; Cloud VDS M $74.5/h; Cloud VDS XXL $209/h
- Core VPS 24-month introductory promotion (USD): Cloud VPS 4 $undefined/h; Cloud VPS 6 $undefined/h; Cloud VPS 8 $undefined/h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, open internet, static egress IP, private networking, GPU, Windows, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## CoreWeave (coreweave)
https://www.coreweave.com/pricing · gpu-cloud · isolation: container
- B200 HGX, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- H100 HGX, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- H200 HGX, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- L40S, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, Docker inside, preinstalled agents, SSH, public IPv4, HTTPS preview URLs, egress allowlist, static egress IP, private networking, SOC 2, SSO, GPU, arm64, Python, Node.js, live fork (no pause), EU data residency, MCP server, snapshots on demand, agent harness API, inbound access rules, extra volumes, shared volumes

## CreateOS Sandbox (NodeOps) (createos)
https://createos.sh/products/sandbox · agent-sandbox · isolation: firecracker
- Sandbox shapes, per-second at list rates (paid from plan credits / top-ups): s-0.25vcpu-512mb $0.014836/h; s-0.5vcpu-1gb $0.029672/h; s-1vcpu-256mb $0.039061/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser + desktop control, SSH, HTTPS preview URLs, egress allowlist, open internet, private networking, self-hosting, BYOC, SSO, Python, Node.js, Go, Rust, wake on request, memory fork, webhooks, EU data residency, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), shared volumes

## Crusoe (crusoe)
https://www.crusoe.ai/cloud/pricing · gpu-cloud · isolation: vm
- H100 SXM, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- H100 SXM, spot (×8 node): $0/vCPU-h, $0/GiB-h
- H200 SXM, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- H200 SXM, spot (×8 node): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Python, audit logs, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Cua Fleet (cua)
https://cua.ai/pricing · agent-sandbox · isolation: vm
- Fleet pool, Ubuntu 24.04 desktop VM (per warm replica-hour): $0.044625/vCPU-h, $0.0223125/GiB-h
- Fleet pool, Windows Server 2022 desktop VM: $0.044625/vCPU-h, $0.0223125/GiB-h
- Cloud macOS (Apple Silicon, Lume/VZ), price not published
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, SSH, HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, arm64, Windows, macOS, Python, Node.js, MCP server, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Cua Fleet (Windows Server) (cua-windows)
https://cua.ai/pricing · windows · isolation: vm
- Cua Fleet usage-based, Windows Server 2022 image: $0.044625/vCPU-h, $0.0223125/GiB-h
- Cua Cloud macOS (separate product): Cloud macOS $null/h
Features: ≥24 h sessions, custom image (Docker or snapshot), full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, Windows, Python, Node.js, MCP server, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Cube Computer (cube)
https://cube.computer/ · dev-env · isolation: vm
- Per-second VM sizes (50 GB disk): 2 CPU 4 GB / 50 GB disk $0.024735/h; 2 CPU 8 GB / 50 GB disk $0.033877/h; 4 CPU 8 GB / 50 GB disk $0.056026/h
Features: pause/resume, persistent disk, ≥24 h sessions, full VM (own kernel), Docker inside, preinstalled agents, SSH, Python, Node.js, gVisor or VM (no shared kernel), firewall inside (nftables)

## Cyberdesk (cyberdesk)
https://www.cyberdesk.io · windows · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, desktop GUI, computer-use API, open internet, BYOC, SOC 2, HIPAA, Windows, Python, Node.js, webhooks, hosted agent API (their own agent), firewall inside (nftables)

## Dagger (dagger)
https://dagger.io · dev-env · isolation: container
- Native product; see regime notes
- Cloud Engines (hosted, early access)
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, self-hosting, open source, SOC 2, SSO, arm64, Python, Node.js, MCP server, snapshots on demand, hosted agent API (their own agent), inbound access rules, extra volumes, shared volumes

## Datalayer Runtimes (datalayer)
https://datalayer.io/pricing · dev-env · isolation: container
- On-demand: 1-CPU $0.3/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, code interpreter, egress allowlist, self-hosting, BYOC, SOC 2, SSO, Python, live fork (no pause), audit logs, spend limits, MCP server, snapshots on demand, hosted agent API (their own agent), extra volumes, shared volumes

## Daytona (daytona)
https://www.daytona.io/pricing · agent-sandbox · isolation: container
- Sandbox, container class (default): $0.0504/vCPU-h, $0.0162/GiB-h
- Linux VM sandbox (same rates; KVM / nested virt per-org feature): $0.0504/vCPU-h, $0.0162/GiB-h
- GPU sandbox, on-demand: $0.0504/vCPU-h, $0.0162/GiB-h
- GPU sandbox, spot / preemptible: $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, Docker inside, nested virtualization, root, browser, desktop GUI, computer-use API, code interpreter, desktop control API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, private networking, BYOC, SOC 2, HIPAA, SSO, GPU, Windows, Python, Node.js, credential injection, live resize, memory fork, webhooks, audit logs, EU data residency, scoped API keys, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Daytona (Windows sandboxes) (daytona-windows)
https://www.daytona.io/pricing · windows · isolation: vm
- Windows VM sandbox (presets small/medium/large): $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), root, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, private networking, BYOC, SOC 2, HIPAA, SSO, Windows, live resize, live fork (no pause), memory fork, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, secret proxy, secret proxy for any API

## Declaw (declaw)
https://docs.declaw.ai/platform/billing · agent-sandbox · isolation: firecracker
- Wallet-metered Firecracker sandbox (allocated vCPU + RAM + overlay disk): $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, Python, Node.js, Go, credential injection, HTTP method/path egress rules, live fork (no pause), memory fork, audit logs, EU data residency, spend limits, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Dedalus Labs (dedalus-labs)
https://www.dedaluslabs.ai/pricing · agent-sandbox · isolation: vm
- On-demand: $0.04536/vCPU-h, $0.01458/GiB-h
Features: pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, full VM (own kernel), Docker inside, root, systemd, SSH, BYOC, SSO, Python, wake on request, live resize, audit logs, gVisor or VM (no shared kernel), firewall inside (nftables)

## Deno Sandbox (deno-sandbox)
https://deno.com/deploy/pricing · agent-sandbox · isolation: firecracker
- Sandbox microVM (2 vCPU fixed), active CPU + allocated memory: $0.1/vCPU-h, $0.025/GiB-h
Features: snapshots, persistent disk, volumes, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), root, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, SOC 2, HIPAA, credential injection, EU data residency, spend limits, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes

## Depot (GitHub Actions runners) (depot)
https://depot.dev/pricing · macos · isolation: vm
- macOS runners (M4 / M2), per minute: depot-macos-26 / 15 / 14 (8 CPU / 24 GB / 400 GB) $4.8/h
- Windows Server 2025/2022 runners, per minute: depot-windows-2025 (2 CPU / 8 GB / 100 GB) $0.48/h; depot-windows-2025-4 (4 CPU / 16 GB / 130 GB) $0.96/h; depot-windows-2025-8 (8 CPU / 32 GB / 150 GB) $1.92/h
- Depot GitHub Actions runners, Linux x64: depot-ubuntu-24.04 $0.36/h; depot-ubuntu-24.04-4 $0.72/h; depot-ubuntu-24.04-8 $1.44/h
- Depot GitHub Actions runners, Linux arm64: depot-ubuntu-24.04-arm $0.36/h; depot-ubuntu-24.04-arm-4 $0.72/h; depot-ubuntu-24.04-arm-8 $1.44/h
Features: snapshots, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, browser, preinstalled agents, SSH, egress allowlist, open internet, static egress IP, private networking, BYOC, SOC 2, SSO, GPU, arm64, Windows, macOS, Python, Node.js, EU data residency, spend limits, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## Dexto (dexto)
https://www.dexto.ai/docs/models/ · agent-sandbox · isolation: not published
- Native product; see regime notes
Features: pause/resume, persistent disk, ≥24 h sessions, browser, desktop GUI, BYOC, scheduled runs (cron), spend limits, MCP server, hosted agent API (their own agent)

## DigitalOcean Droplets (digitalocean)
https://www.digitalocean.com/pricing/droplets · hyperscaler · isolation: vm
- Basic Droplet, Regular CPU (shared vCPU), per-second with 672 h monthly cap: s-1vcpu-512mb-10gb (Basic Regular, shared) $0.00595/h; s-1vcpu-1gb (Basic Regular, shared) $0.00893/h; s-1vcpu-2gb (Basic Regular, shared) $0.01786/h
- Basic Droplet, Premium AMD / Intel (shared vCPU, NVMe), per-second with 672 h cap: s-1vcpu-1gb-amd (Basic Premium AMD, shared) $0.01042/h; s-1vcpu-2gb-amd (Basic Premium AMD, shared) $0.02083/h; s-2vcpu-2gb-amd (Basic Premium AMD, shared) $0.03125/h
- Dedicated vCPU (General Purpose / CPU-Optimized / Memory-Optimized / Storage-Optimized, Regular), per-second with 672 h cap: g-2vcpu-8gb (General Purpose Regular, dedicated) $0.09375/h; g-4vcpu-16gb (General Purpose Regular, dedicated) $0.1875/h; g-8vcpu-32gb (General Purpose Regular, dedicated) $0.375/h
- Dedicated vCPU Premium Intel (General Purpose / CPU-Optimized / Memory-Optimized / Storage-Optimized), per-second with 672 h cap: g-2vcpu-8gb-intel (General Purpose Premium Intel, dedicated) $0.1131/h; g-4vcpu-16gb-intel (General Purpose Premium Intel, dedicated) $0.2247/h; g-8vcpu-32gb-intel (General Purpose Premium Intel, dedicated) $0.4494/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, live fork (no pause), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## DigitalOcean GPU Droplets (digitalocean-gpu)
https://www.digitalocean.com/pricing/gpu-droplets · gpu-cloud · isolation: vm
- MI325X, on-demand (×1): $0/vCPU-h, $0/GiB-h
- MI325X, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- MI300X, on-demand (×1): $0/vCPU-h, $0/GiB-h
- MI300X, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Python, live fork (no pause), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Docker Cloud Sandboxes (docker-cloud-sandboxes)
https://www.docker.com/blog/introducing-cloud-sandboxes-start-on-your-laptop-finish-in-the-cloud/ · agent-sandbox · isolation: vm
- Pay-as-you-go fixed sizes (micro..xl), per second: micro $0.07/h; small $0.14/h; medium $0.28/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, SSO, arm64, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes

## Dockup (dockup)
https://getdockup.com/ · paas · isolation: container
- Native service
Features: ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, HTTPS preview URLs, egress allowlist, private networking, BYOC, EU data residency, inbound access rules

## E2B (e2b)
https://e2b.dev/pricing · agent-sandbox · isolation: firecracker
- On-demand (per-second, allocated): $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, systemd, browser, desktop GUI, computer-use API, code interpreter, desktop control API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, open source, SOC 2, HIPAA, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, webhooks, EU data residency, spend limits, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## E2E Networks (e2e-networks)
https://www.e2enetworks.com/pricing · hyperscaler · isolation: vm
- C3 compute-intensive: C3.8GB $0.049/h; C3.16GB $0.096/h; C3.32GB $0.194/h
- Dedicated CPU: SDC3-4.30GB $0.099/h; SDC3-8.60GB $0.189/h; SDC3-12.90GB $0.283/h
- M3 memory-intensive: M3.32GB $0.08/h; M3.48GB $0.161/h; M3.96GB $0.321/h
- E1 shared-core Linux: E1LC-2.6GB $0.044/h; E1LC-4.12GB $0.089/h; E1LC-6.18GB $0.134/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, SOC 2, SSO, GPU, Windows, Python, Node.js, live fork (no pause), snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Ellipsis (ellipsis)
https://www.ellipsis.dev/pricing · agent-sandbox · isolation: qemu-kvm
- On-demand: $0.141912/vCPU-h, $0.024192/GiB-h
- Own AWS VPC
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, preinstalled agents, self-hosting, BYOC, SOC 2, Python, Node.js, wake on request, webhooks, scheduled runs (cron), audit logs, spend limits, automatic snapshots, agent harness API, gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy

## Eventual (eventual)
https://www.eventual.ai/ · paas · isolation: not published
- Daft open-source self-hosted engine
- Eventual robotics data platform
Features: ≥24 h sessions, self-hosting, BYOC, open source, GPU, arm64, Python

## exe.dev (exe-dev)
https://exe.dev/pricing · dev-env · isolation: vm
- Work pool (current; reserved pool billed hourly 24/7, shared by all VMs)
- Personal pool (current; 2/4/8/16 vCPU at $15/$35/$75/$155)
- Standalone / sandbox VMs (current; per VM, own reserved capacity): $0.0525/vCPU-h, $0/GiB-h
- Usage / Cloud Pool (sales-only, per second, peak-hourly CPU + active memory): $0.05/vCPU-h, $0.016/GiB-h
Features: fork/clone, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, systemd, browser, preinstalled agents, SSH, HTTPS preview URLs, custom domains, open internet, private networking, BYOC, SSO, Python, Go, credential injection, live resize, live fork (no pause), EU data residency, scoped API keys, MCP server, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Exoscale (exoscale)
https://www.exoscale.com/pricing/ · hyperscaler · isolation: vm
- Standard: standard-micro $0.00729/h; standard-tiny $0.01458/h; standard-small $0.02333/h
- Cpu: cpu-extra_large $0.23333/h; cpu-huge $0.46666/h; cpu-mega $0.93333/h
- Memory: memory-extra_large $0.14/h; memory-huge $0.28/h; memory-mega $0.56/h
- Storage optimized: storage-extra_large $0.14/h; storage-huge $0.28/h; storage-mega $0.56/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Windows, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Expanse (expanse)
https://expanse.sh · gpu-cloud · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, self-hosting, BYOC, SOC 2, EU data residency

## Expo (expo)
https://expo.dev · macos · isolation: vm
- iOS build, medium worker: iOS medium (5 performance cores / 20 GiB, Mac mini) $0/h
- iOS build, large worker: iOS large (10 performance cores / 40 GiB, Mac mini) $0/h
- Android build, medium worker: Android medium (4 vCPU / 16 GB, GCE n2/c3d-standard-4) $0/h
- Android build, large worker: Android large (8 vCPU / 32 GB, GCE n2/c3d-standard-8) $0/h
Features: full VM (own kernel), nested virtualization, open internet, static egress IP, SOC 2, SSO, arm64, macOS, Node.js, scheduled runs (cron), MCP server, gVisor or VM (no shared kernel), firewall inside (nftables)

## fal (fal)
https://fal.ai/pricing · gpu-cloud · isolation: container
- B300, serverless (×1): $0/vCPU-h, $0/GiB-h
- B300, reserved
- B200, serverless (×1): $0/vCPU-h, $0/GiB-h
- B200, reserved
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, SSH, public IPv4, HTTPS preview URLs, SOC 2, SSO, GPU, Python, Node.js, wake on request, MCP server, hosted agent API (their own agent), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Ferr (ferr)
https://ferr.dev/ · browser · isolation: container
- Cloud browser overage unpublished: Cloud browser overage unpublished $null/h
Features: persistent disk, idle auto-stop, ≥24 h sessions, browser, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, residential IPs, open internet, static egress IP, self-hosting, open source, Python, Node.js

## Firecrawl Interact / Browser Sandbox (firecrawl)
https://www.firecrawl.dev/pricing · browser · isolation: container
- Hobby code-only Interact: Hobby $0.6/h
- Standard code-only Interact: Standard $0.3/h
- Growth code-only Interact: Growth $0.24/h
- Scale code-only Interact: Scale $0.12/h
Features: idle auto-stop, browser, browser automation API, anti-bot stealth, egress allowlist, open internet, static egress IP, SOC 2, SSO, Python, Node.js, MCP server, hosted agent API (their own agent), inbound access rules

## Flow Swiss Mac Bare Metal (flow-swiss-mac)
https://doc.flow.swiss/platform/pricing/mac-bare-metal · macos · isolation: dedicated-host
- Whole Mac, hourly with 24 h minimum (CHF converted to USD): macmini.m1.8-16-512 (M1) $0.326295/h; macmini.2018.6-16-256 (Intel i7) $0.326295/h; macmini.m2.8-24-512 (M2) $0.41089/h
- Mac split into 2 macOS VMs (Apple SLA cap), per-VM price: 1/2 of macmini.m1.8-16-512 (M1) $0.163148/h; 1/2 of macmini.m2.8-24-512 (M2) $0.205445/h; 1/2 of macmini.m2-pro.12-32-4096 (M2 Pro) $0.374635/h
- CI Engine (managed ephemeral macOS CI runners on M4 Pro), monthly per concurrent job: Essential gen1-3c.8g $null/h; Professional gen1-6c.16g $null/h
Features: persistent disk, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, arm64, macOS, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## FlowDeploy (flowdeploy)
https://flowdeploy.com · gpu-cloud · isolation: not published · discontinued

Features: ≥24 h sessions

## Fluidstack (fluidstack)
https://fluidstack.io/ · gpu-cloud · isolation: dedicated-host
- H100, reserved
- H200, reserved
- B200, reserved
- GB200, reserved
Features: pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), full VM (own kernel), Docker inside, nested virtualization, SSH, SSO, GPU, Python, Node.js, gVisor or VM (no shared kernel), firewall inside (nftables)

## Fly.io Machines (fly-machines)
https://fly.io/pricing · paas · isolation: firecracker
- Performance CPU (dedicated), cheapest us region: $0.02916/vCPU-h, $0.0083376/GiB-h
- Performance CPU (dedicated), cheapest eu region: $0.02916/vCPU-h, $0.0083376/GiB-h
- Performance CPU (dedicated), cheapest asia region: $0.02916/vCPU-h, $0.0083376/GiB-h
- Performance CPU (dedicated), cheapest other region: $0.02916/vCPU-h, $0.0083376/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, open source, SOC 2, HIPAA, Python, Node.js, wake on request, live fork (no pause), EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Freestyle (freestyle)
https://www.freestyle.sh/pricing · agent-sandbox · isolation: bare-metal-vm
- On-demand (allocated, per second): $0.04032/vCPU-h, $0.0129/GiB-h
- Enterprise active-CPU pricing
- Enterprise node-based (dedicated capacity, flat rate)
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, audit logs, scoped API keys, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Gcore Cloud / Functions / GPU (gcore)
https://gcore.com/cloud/virtual-machines · hyperscaler · isolation: vm
- Virtual machine: g0-standard-1-2 $0.0091/h; g0-standard-2-4 $0.01934/h; g0-standard-4-8 $0.03869/h
- Functions
- GPU instance: Virtual GPU 1xH100 (24 vCPU / 232 GB, 1250 GB) $0/h
- Windows dedicated VM families
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, static egress IP, private networking, GPU, arm64, Windows, Python, Node.js, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Gemini API code execution (gemini-code-execution)
https://ai.google.dev/gemini-api/docs/code-execution · agent-sandbox · isolation: container
- Built-in code execution tool (token-billed only): code execution (size not published) $0/h
Features: idle auto-stop, code interpreter, Python, spend limits

## GitHub Actions hosted runners (github-actions)
https://docs.github.com/en/billing/reference/actions-runner-pricing · macos · isolation: vm
- macOS runners, per minute: macOS 3-core M1 (standard, 7 GB) $3.72/h; macOS 5-core M2 Pro (xlarge, 14 GB) $6.12/h; macOS 12-core Intel (large, 30 GB) $4.62/h
- Linux x64 standard runner (2-core), per minute: Linux 2-core standard (actions_linux) $0.36/h
- Windows x64 runners, per minute: Windows 2-core (standard) $0.6/h; Windows 4-core (larger) $1.32/h; Windows 8-core $2.52/h
- Linux x64 larger runners, per minute: Linux 4-core $0.72/h; Linux 8-core $1.32/h; Linux 16-core $2.52/h
Features: snapshots, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, browser, preinstalled agents, egress allowlist, open internet, static egress IP, private networking, self-hosting, open source, SOC 2, SSO, GPU, arm64, Windows, macOS, Python, Node.js, Go, Java, webhooks, audit logs, EU data residency, spend limits, MCP server, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## GitHub Codespaces (github-codespaces)
https://docs.github.com/en/billing/concepts/product-billing/github-codespaces · dev-env · isolation: vm
- Pay-as-you-go machine types (per active hour): 2-core $0.18/h; 4-core $0.36/h; 8-core $0.72/h
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, SSH, HTTPS preview URLs, open internet, SOC 2, SSO, Python, Node.js, Go, Java, audit logs, EU data residency, spend limits, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## GitHub Copilot cloud sandboxes (github-copilot-cloud-sandboxes)
https://docs.github.com/en/billing/concepts/product-billing/cloud-and-local-sandboxes · hyperscaler · isolation: firecracker
- Cloud sandbox (copilot --cloud), per-second meters: $0.0864/vCPU-h, $0.0108/GiB-h
Features: snapshots, memory snapshots, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), root, preinstalled agents, egress allowlist, open internet, SOC 2, SSO, Python, Node.js, HTTP method/path egress rules, spend limits, MCP server, automatic snapshots, agent harness API, gVisor or VM (no shared kernel), firewall inside (nftables)

## GitLab.com hosted runners (gitlab-runners)
https://docs.gitlab.com/ci/pipelines/compute_minutes/ · macos · isolation: vm
- macOS hosted runners (beta), compute minutes x cost factor: saas-macos-medium-m1 (4 vCPU / 8 GB, factor 6) $3.6/h; saas-macos-large-m2pro (6 vCPU / 16 GB, factor 12) $7.2/h
- Windows hosted runners (beta), compute minutes x cost factor: saas-windows-medium-amd64 (2 vCPU / 7.5 GB, factor 1) $0.6/h
- GitLab.com hosted Linux x86 (small/medium): saas-linux-small-amd64 $0.6/h; saas-linux-medium-amd64 $1.2/h
- GitLab.com hosted Linux x86 large+ (Premium/Ultimate): saas-linux-large-amd64 $1.8/h; saas-linux-xlarge-amd64 $3.6/h; saas-linux-2xlarge-amd64 $7.2/h
Features: idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, preinstalled agents, self-hosting, open source, SOC 2, SSO, GPU, arm64, Windows, macOS, Python, Node.js, webhooks, scheduled runs (cron), audit logs, agent harness API, gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy

## GKE Agent Sandbox (gke-agent-sandbox)
https://cloud.google.com/kubernetes-engine/docs/concepts/agent-sandbox · hyperscaler · isolation: gvisor
- GKE Autopilot sandbox pods: $0.0445/vCPU-h, $0.0049225/GiB-h
- GKE Autopilot Spot sandbox pods: $0.0133/vCPU-h, $0.0014767/GiB-h
- GKE Standard sandbox nodes
- Autopilot default — Compute Flexible CUD 1 year: $0.03204/vCPU-h, $0.0035442/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, HTTPS preview URLs, egress allowlist, open internet, static egress IP, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, arm64, Python, live resize, live fork (no pause), memory fork, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## Gologin Cloud Browser (gologin)
https://gologin.com/cloud-browser/ · browser · isolation: container
- Professional browser: Professional $0.09/h
- Business browser: Business $0.08/h
- Enterprise browser: Enterprise $0.07/h
Features: snapshots, fork/clone, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, anti-bot stealth, residential IPs, open internet, static egress IP, SOC 2, SSO, Python, Node.js, wake on request, audit logs, MCP server, automatic snapshots

## Google Agent Platform Sandbox (Code Execution / Shell / Computer Use) (vertex-computer-use)
https://cloud.google.com/products/gemini-enterprise-agent-platform/pricing · agent-sandbox · isolation: container
- Code Execution / Shell sandbox (API exec, same Agent Compute + Agent Memory meter): $0.085/vCPU-h, $0.009/GiB-h: default (MACHINE_CONFIG_UNSPECIFIED) $0.1835/h; MACHINE_CONFIG_VCPU4_RAM4GIB $0.376/h
- Computer Use sandbox (managed browser, billed as Agent Compute + Agent Memory): $0.085/vCPU-h, $0.009/GiB-h
- Computer Use / Shell sandbox on Gemini Enterprise Flexible Savings Plan (1 year): $0.0765/vCPU-h, $0.0081/GiB-h
- Computer Use / Shell sandbox on Gemini Enterprise Flexible Savings Plan (3 year): $0.068/vCPU-h, $0.0072/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, browser, computer-use API, code interpreter, browser automation API, desktop control API, browser + desktop control, egress allowlist, static egress IP, private networking, SOC 2, HIPAA, SSO, arm64, Python, Node.js, HTTP method/path egress rules, live fork (no pause), memory fork, EU data residency, snapshots on demand, agent harness API, inbound access rules

## Google Agent Runtime / Agent Engine (google-agent-engine)
https://cloud.google.com/products/gemini-enterprise-agent-platform/pricing · hyperscaler · isolation: gvisor
- Agent Runtime allocation while processing turns: $0.085/vCPU-h, $0.009/GiB-h
- Runtime 1-year Flexible Savings Plan: $0.0765/vCPU-h, $0.0081/GiB-h
- Runtime 3-year Flexible Savings Plan: $0.068/vCPU-h, $0.0072/GiB-h
- Code Execution Sandbox (separate from agent Runtime): default $0.1835/h; VCPU4_RAM4GIB $0.376/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, egress allowlist, static egress IP, private networking, SOC 2, HIPAA, SSO, arm64, Python, Node.js, HTTP method/path egress rules, live fork (no pause), memory fork, EU data residency, automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, extra volumes

## Google Cloud Build (google-cloud-build)
https://cloud.google.com/build/pricing · paas · isolation: vm
- Default pool e2-standard-2: e2-standard-2 $0.36/h
- Default pool (e2-medium / e2-highcpu): e2-medium $0.18/h; e2-highcpu-8 $0.936/h; e2-highcpu-32 $3.744/h
- Private pool E2 standard: e2-standard-2 $0.36/h; e2-standard-4 $0.72/h; e2-standard-8 $1.44/h
- Private pool C3 high-CPU (Iowa/reference regions): c3-highcpu-4 $0.83664/h; c3-highcpu-8 $1.67328/h; c3-highcpu-22 $4.60152/h
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, root, egress allowlist, open internet, static egress IP, SOC 2, HIPAA, SSO, Python, Node.js, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables)

## Google Cloud Run (google-cloud-run)
https://cloud.google.com/run/pricing · hyperscaler · isolation: gvisor
- Services, instance-based billing ("CPU always allocated"), Tier 1: $0.0648/vCPU-h, $0.0072/GiB-h
- Services, request-based billing (default), Tier 1: $0.0864/vCPU-h, $0.009/GiB-h
- Services/Jobs, instance-based billing, second-gen execution environment (microVM VM sandbox), Tier 1: $0.0648/vCPU-h, $0.0072/GiB-h
- Cloud Run CUD 1y/3y, instance-based: $0.053784/vCPU-h, $0.005976/GiB-h
Features: volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, root, HTTPS preview URLs, custom domains, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, webhooks, audit logs, EU data residency, spend limits, MCP server, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## Google Colab (google-colab)
https://developers.google.com/colab · dev-env · isolation: vm
- On-demand: 
- Colab Enterprise (GCP, us-central1): $0.026173908/vCPU-h, $0.003508236/GiB-h
Features: pause/resume, persistent disk, volumes, idle auto-stop, full VM (own kernel), code interpreter, HTTPS preview URLs, egress allowlist, open internet, SOC 2, HIPAA, SSO, GPU, Python, EU data residency, MCP server, gVisor or VM (no shared kernel), inbound access rules, secret proxy, extra volumes

## Google Compute Engine (Linux VMs) (gcp-compute)
https://cloud.google.com/products/compute/pricing/general-purpose · hyperscaler · isolation: vm
- On-demand E2 / N2D / C3D (us-central1): c3d-standard-4 $0.181596/h; c3d-standard-8 $0.363192/h; c3d-standard-16 $0.726384/h
- On-demand Tau T2A Arm (us-central1): t2a-standard-1 $0.0385/h; t2a-standard-2 $0.077/h; t2a-standard-4 $0.154/h
- Spot E2 / N2D / C3D (us-central1): c3d-standard-4 $0.048728/h; c3d-standard-8 $0.097456/h; c3d-standard-16 $0.194912/h
- Spot Tau T2A Arm (us-central1): t2a-standard-1 $0.0231/h; t2a-standard-2 $0.0462/h; t2a-standard-4 $0.0924/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, HTTP method/path egress rules, live fork (no pause), audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Google Compute Engine (Windows Server) (gcp-windows)
https://cloud.google.com/compute/disks-image-pricing#windows_server_pricing · windows · isolation: vm
- E2 custom machine + Windows Server licence (us-central1): $0.02290217/vCPU-h, $0.003069707/GiB-h
- E2 predefined machine types + Windows licence (us-central1): e2-highcpu-2 $0.04947024/h; e2-standard-2 $0.06701142/h; e2-highmem-2 $0.09039966/h
- E2 custom + Windows licence (europe-west1): $0.02519304/vCPU-h, $0.003376895/GiB-h
- E2 shared-core + Windows licence (us-central1): e2-small $0.016752855/h; e2-medium $0.03350571/h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, browser, desktop GUI, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, HTTP method/path egress rules, audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Google Compute Engine GPU VMs (gcp-compute-gpu)
https://cloud.google.com/products/compute/pricing/accelerator-optimized · gpu-cloud · isolation: vm
- a2-highgpu-1g, on-demand (×1): $0/vCPU-h, $0/GiB-h
- a2-ultragpu-1g, on-demand (×1): $0/vCPU-h, $0/GiB-h
- a3-highgpu-8g, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- a3-megagpu-8g, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Windows, Python, HTTP method/path egress rules, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Green Mini host (greenmini)
https://portal.greenmini.host/checkout/order · macos · isolation: dedicated-host
- Dedicated Mac (portal list price per month): Mac mini M2 8-core, 8 GB, 256 GB (EUR 59/mo) $0.094562/h; Mac mini M4 10-core, 16 GB, 256 GB ("from" price) (EUR 74/mo) $0.118603/h; Mac mini M4 10-core, 16 GB, 512 GB (EUR 87/mo) $0.139438/h
- No-commitment monthly M4: Mac mini M4 10-core, 16 GB, 256 GB (€84 monthly) $undefined/h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, open internet, static egress IP, private networking, arm64, macOS, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## H Company / Surfer / H Agents API (h-company)
https://www.hcompany.ai/pricing · browser · isolation: not published
- Managed H Agents browser: Managed H Agents browser $null/h
Features: pause/resume, idle auto-stop, ≥24 h sessions, browser, computer-use API, browser automation API, residential IPs, open internet, static egress IP, self-hosting, open source, SOC 2, webhooks, scheduled runs (cron), EU data residency, MCP server, hosted agent API (their own agent)

## Halluminate (halluminate)
https://halluminate.ai/ · agent-sandbox · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents

## Hatchet (hatchet-run)
https://hatchet.run · paas · isolation: container
- Native product; see regime notes
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, webhooks, scheduled runs (cron), audit logs, EU data residency, MCP server

## Heroic Labs (heroic-labs)
http://heroiclabs.com · paas · isolation: dedicated-vm
- Native product; see regime notes
Features: persistent disk, ≥24 h sessions, full VM (own kernel), HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, SSO, Node.js, EU data residency, gVisor or VM (no shared kernel)

## Heroku (heroku)
https://www.heroku.com/pricing/ · paas · isolation: container
- Cedar Common Runtime dynos (Basic/Standard shared; Performance dedicated): Basic $0.00972/h; Standard-1X $0.03472/h; Standard-2X $0.06944/h
- Eco dynos: $5 flat for a 1,000 dyno-hour pool (personal apps only)
- Cedar Private Space dynos (dedicated; requires a Private Space): Private-S $0.17361/h; Private-M $0.34722/h; Private-L $0.69444/h
- Cedar Shield Private Space dynos (HIPAA/PCI): Shield-S $0.20833/h; Shield-M $0.41667/h; Shield-L $0.83333/h
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, SSH, HTTPS preview URLs, custom domains, static egress IP, private networking, SOC 2, HIPAA, SSO, arm64, Python, Node.js, Go, Java, wake on request, webhooks, EU data residency, MCP server, inbound access rules

## Hetzner Cloud (hetzner-cloud)
https://docs.hetzner.com/general/infrastructure-and-availability/price-adjustment/ · hyperscaler · isolation: vm
- Shared vCPU, Regular Performance (CPX, AMD), EU (NBG1/HEL1): CPX22 $0.0368/h; CPX32 $0.0673/h; CPX42 $0.1314/h
- Shared vCPU, Regular Performance (CPX, AMD), US (ASH1/HIL1): CPX11 $0.0328/h; CPX21 $0.0601/h; CPX31 $0.1178/h
- Shared vCPU, Regular Performance (CPX, AMD), Singapore (SIN1): CPX12 $0.0288/h; CPX22 $0.0497/h; CPX32 $0.0929/h
- Dedicated vCPU, General Purpose (CCX, AMD), EU (NBG1/HEL1): CCX13 $0.0809/h; CCX23 $0.1626/h; CCX33 $0.2612/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, arm64, Python, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Hetzner Dedicated AX / EX (hetzner-dedicated)
https://www.hetzner.com/dedicated-rootserver/ · hyperscaler · isolation: dedicated-host
- AX physical server hourly with monthly cap: AX42 (+$59 one-time setup) $0.1877/h; AX102 (+$149 one-time setup) $0.4841/h; AX162 (+$359 one-time setup) $1.1572/h
- EX physical server hourly with monthly cap: EX44 (setup fee unverified) $0.11549999999999999/h; EX63 (+$84 one-time setup) $0.2758/h; EX131 (+$334 one-time setup) $1.0691000000000002/h
- Server Auction (refurbished, Dutch auction, no setup fee): SB i7-7700 (4c/8t) 64GB 2x4TB HDD $0.109/h; SB i7-6700 (4c/8t) 64GB 2x512GB NVMe $0.1106/h; SB Ryzen 5 3600 (6c/12t) 64GB 2x2TB HDD $0.1154/h
- GEX GPU dedicated server hourly with monthly cap: GEX45 RTX PRO 4000 Blackwell SFF (+$249 setup) $0.396/h; GEX131 RTX PRO 6000 Blackwell Max-Q (+$699 setup) $2.2389/h
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, GPU, arm64, Windows, Python, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Hetzner GPU servers (hetzner-gpu)
https://www.hetzner.com/dedicated-rootserver/matrix-gpu/ · gpu-cloud · isolation: dedicated-host
- GEX45 Blackwell SFF 24GB, on-demand (×1): GEX45 $0.399/h
- GEX131 Blackwell Max-Q 96GB, on-demand (×1): GEX131-1 $2.2419/h
- GEX131-2 Blackwell Max-Q 96GB: GEX131-2 $null/h
- GEX131-3 Blackwell Max-Q 96GB: GEX131-3 $null/h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, GPU, Windows, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Hoplite (hoplite)
https://hoplite.sh · agent-sandbox · isolation: gvisor
- Native service
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, Docker inside, browser, preinstalled agents, HTTPS preview URLs, Python, Node.js, wake on request, scheduled runs (cron), spend limits, MCP server, automatic snapshots, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules

## Hopx (hopx)
https://hopx.ai/pricing · agent-sandbox · isolation: firecracker
- Pay-per-use (per second, allocated, template-fixed size): $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, HTTPS preview URLs, egress allowlist, open internet, SOC 2, SSO, Python, Node.js, Go, webhooks, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables)

## Hostinger VPS (hostinger-vps)
https://www.hostinger.com/vps-hosting · hyperscaler · isolation: vm
- Regular price, 24-month prepaid (amortized monthly): KVM 1 $11.99/h; KVM 2 $14.99/h; KVM 4 $28.99/h
- Intro price, 24-month prepaid (renews at regular price, e.g. KVM 4 $12.99 -> $28.99): KVM 1 $6.49/h; KVM 2 $8.99/h; KVM 4 $12.99/h
- 1-month term, no commitment (auto-renews monthly): KVM 1 $19.49/h; KVM 2 $24.49/h; KVM 4 $42.99/h
- Regular price, 12-month prepaid: KVM 1 $12.99/h; KVM 2 $16.99/h; KVM 4 $30.99/h
Features: snapshots, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, desktop GUI, SSH, public IPv4, custom domains, raw TCP inbound, open internet, static egress IP, SSO, GPU, Python, Node.js, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## HostMyApple (hostmyapple)
https://hostmyapple.com/mac-vps-hosting · macos · isolation: vm
- macOS Tahoe cloud VPS (virtualized macOS), monthly: Cloud Lite: 4 vCPU, 3 GB, 60 GB SSD $0.034233/h; Cloud Basic: 4 vCPU, 4 GB, 80 GB SSD $0.047932/h; Cloud Pro: 4 vCPU, 5 GB, 100 GB SSD $0.06163/h
- Dedicated Mac mini, monthly: Mac mini M6, 16 GB, 256 GB (11 available) $0.191767/h; Mac mini M6, 32 GB, 256 GB (2 available) $0.246562/h; Mac Studio M5 Max 4.6 GHz, 64 GB, 512 GB (portal only, 1 available) $0.479452/h
Features: persistent disk, ≥24 h sessions, full VM (own kernel), root, browser, desktop GUI, SSH, public IPv4, open internet, static egress IP, arm64, macOS, Python, Node.js, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables)

## Huawei Cloud AgentArts (huawei-cloud)
https://support.huaweicloud.com/price-agentarts/agentarts-price-pdf.pdf · agent-sandbox · isolation: container
- Base CU tariff, mainland China (public beta): code interpreter/browser/runtime: $0.165531779128/vCPU-h, $0.016553177913/GiB-h
- ECS pay-per-use, AP-Singapore (international site, USD): general computing S6 / X1 / C9: s6.medium.2 $0.033/h; s6.large.2 $0.066/h; s6.large.4 $0.08/h
- ECS pay-per-use, LA-Mexico City2 (cheapest international region): S6 / X1: s6.medium.2 $0.027/h; s6.large.2 $0.054/h; s6.large.4 $0.06/h
- ECS yearly/monthly subscription, AP-Singapore: s6.large.2 $35.92/h; s6.xlarge.2 $71.83/h; s6.2xlarge.2 $143.66/h
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, browser, computer-use API, code interpreter, browser + desktop control, HTTPS preview URLs, egress allowlist, open internet, private networking, SSO, arm64, Python, Node.js, wake on request, agent harness API, hosted agent API (their own agent), inbound access rules, secret proxy, extra volumes, shared volumes

## HUD (hud)
https://www.hud.ai/ · agent-sandbox · isolation: vm
- Native service: Micro (t3.small) $null/h; Small (t3.medium) $null/h; Default (t3.large) $null/h
- HUD-hosted on Modal (CPU/GPU profiles): cpu=1;memory_mb=8192 $null/h; cpu=2;memory_mb=8192 $null/h; cpu=4;memory_mb=8192 $null/h
Features: ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, egress allowlist, self-hosting, BYOC, open source, SOC 2, GPU, Python, spend limits, MCP server, agent harness API, inbound access rules, secret proxy

## Hugging Face Jobs (hf-sandbox backend) (huggingface-jobs)
https://huggingface.co/docs/hub/en/jobs-pricing · paas · isolation: container
- On-demand: cpu-basic $0.01/h; cpu-upgrade $0.03/h; cpu-xl $1/h
- GPU: Nvidia T4: t4-small $0/h; t4-medium $0.2/h
- GPU: Nvidia L4: l4x1 $0/h
- GPU: Nvidia L40S: l40sx1 $0/h
Features: volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, SSH, HTTPS preview URLs, open internet, SOC 2, HIPAA, SSO, GPU, Python, webhooks, scheduled runs (cron), EU data residency, spend limits, inbound access rules, extra volumes, shared volumes

## HumanLayer (humanlayer)
https://humanlayer.com · agent-sandbox · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, Docker inside, preinstalled agents, private networking, self-hosting, BYOC, SOC 2, SSO, arm64, Windows, macOS, audit logs, zero data retention, agent harness API

## Hyperbeam (hyperbeam)
https://hyperbeam.com/ · browser · isolation: not published
- HD participant-hour: HD participant-hour $0.42/h
Features: idle auto-stop, ≥24 h sessions, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, open internet, Node.js, webhooks, EU data residency, inbound access rules

## Hyperbolic (hyperbolic)
https://www.hyperbolic.ai/marketplace · gpu-cloud · isolation: vm
- H100 SXM · on-demand (marketplace floor): $0/vCPU-h, $0/GiB-h
- H200 · on-demand (marketplace floor): $0/vCPU-h, $0/GiB-h
- B200 · on-demand (marketplace floor): $0/vCPU-h, $0/GiB-h
- Reserved term (prepaid, 7/14/30 days): $0/vCPU-h, $0/GiB-h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, root, SSH, public IPv4, raw TCP inbound, GPU, Python, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Hyperbrowser (hyperbrowser)
https://www.hyperbrowser.ai/pricing · browser · isolation: firecracker
- Cloud browser session ($0.10 per browser-hour, credits): browser session $0.1/h
- Hyperbrowser Sandbox (Firecracker VM), per vCPU-h + GiB-h: $0.045/vCPU-h, $0.015/GiB-h
Features: snapshots, memory snapshots, fork/clone, volumes, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, browser, computer-use API, browser automation API, desktop control API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, static egress IP, HIPAA, Python, Node.js, live fork (no pause), memory fork, EU data residency, MCP server, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Hyperstack (hyperstack)
https://www.hyperstack.cloud/gpu-pricing · gpu-cloud · isolation: vm
- H100 PCIe 80GB (n3-H100x1) on-demand: $0/vCPU-h, $0/GiB-h
- A100 80GB PCIe (n3-A100x1) on-demand: $0/vCPU-h, $0/GiB-h
- RTX Pro 6000 SE (n3-RTX-PRO6000-SEx1) on-demand: $0/vCPU-h, $0/GiB-h
- L40 (n3-L40x1) on-demand: $0/vCPU-h, $0/GiB-h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, egress allowlist, open internet, static egress IP, SOC 2, GPU, Python, Node.js, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Hyrex (hyrex)
https://www.hyrex.io · paas · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, self-hosting, BYOC, open source, Python, Node.js

## IBM Cloud VPC (ibm-cloud-vpc)
https://www.ibm.com/products/virtual-servers/pricing · hyperscaler · isolation: vm
- Gen3/Flex US-South list: bx3d-16x80 $0.907/h; bx3d-24x120 $1.36/h; bx3d-2x10 $0.113/h
- Flex50% baseline: bxf-16x64 $0.5244/h; bxf-2x8 $0.0655/h; bxf-4x16 $0.1311/h
- Catalog spot rate reference: bxf-16x64 $0.192668/h; bxf-24x96 $0.289001/h; bxf-2x8 $0.024083/h
- Reserved VPC capacity: 
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Windows, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## InsForge (InstaCloud) (insforge-instacloud)
https://www.instacloud.com/pricing · paas · isolation: vm
- Native product; see regime notes
- InstaCloud compute (microVM service, usage-billed): $0.027792/vCPU-h, $0.013896/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, SSH, HTTPS preview URLs, custom domains, self-hosting, open source, SOC 2, HIPAA, SSO, wake on request, live fork (no pause), scheduled runs (cron), EU data residency, scoped API keys, MCP server, automatic snapshots, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes

## InstaVM (instavm)
https://instavm.io/pricing · agent-sandbox · isolation: firecracker
- Firecracker microVM, per-second allocated vCPU + RAM: $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, SOC 2, Python, Node.js, credential injection, wake on request, live resize, live fork (no pause), webhooks, audit logs, automatic snapshots, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Intuned (intuned)
https://intunedhq.com/pricing · browser · isolation: vm
- Standard browser machine, Developer plan overage: Standard machine $0.12/h; Large machine $0.3/h; X-Large machine $0.6/h
- Standard browser machine, Startup plan overage: Standard machine $0.1/h; Large machine $0.25/h; X-Large machine $0.5/h
Features: idle auto-stop, full VM (own kernel), browser, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, SSO, Python, Node.js, wake on request, EU data residency, spend limits, hosted agent API (their own agent), gVisor or VM (no shared kernel)

## IONOS Cloud (ionos)
https://docs.ionos.com/cloud/support/general-information/price-list/ionos-cloud-eur-en · hyperscaler · isolation: vm
- Basic Cloud Cubes: XS $0.0079646/h; S $0.0147914/h; M $0.0273072/h
- Compute Engine vCPU: $0.0136536/vCPU-h, $0.0022756/GiB-h
- Cloud GPU VM H200 PCIe: H200-S $3.4134/h; H200-M $6.8268/h; H200-L $13.6536/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, SSO, GPU, Windows, Python, Node.js, live fork (no pause), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Isle (isle)
https://www.tryisle.com/ · agent-sandbox · isolation: not published
- Native product meter; not a general-purpose 4-vCPU/8-GiB VM: 
Features: snapshots, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, desktop GUI, computer-use API, code interpreter, Python, snapshots on demand

## Islo (islo)
https://islo.dev/pricing · agent-sandbox · isolation: vm
- Pay-as-you-go (Team list rates): $0.07/vCPU-h, $0.04/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, private networking, BYOC, SSO, credential injection, HTTP method/path egress rules, wake on request, live fork (no pause), webhooks, scheduled runs (cron), audit logs, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Isorun (isorun)
https://docs.isorun.ai/getting-started/pricing · agent-sandbox · isolation: vm
- MicroVM sandbox, per-second (public beta): $0.025/vCPU-h, $0.015/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, browser + desktop control, HTTPS preview URLs, egress allowlist, open internet, Python, Node.js, credential injection, HTTP method/path egress rules, live fork (no pause), memory fork, audit logs, EU data residency, MCP server, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, extra volumes

## Jamsocket (jamsocket)
https://jamsocket.com · agent-sandbox · isolation: gvisor · discontinued

Features: ≥24 h sessions, self-hosting, open source, automatic snapshots, inbound access rules

## Jarvislabs (jarvislabs)
https://jarvislabs.ai/pricing · gpu-cloud · isolation: vm
- H200 SXM on-demand: $0/vCPU-h, $0/GiB-h
- H100 SXM on-demand: $0/vCPU-h, $0/GiB-h
- RTX Pro 6000 Blackwell on-demand: $0/vCPU-h, $0/GiB-h
- A100 80GB on-demand: $0/vCPU-h, $0/GiB-h
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), full VM (own kernel), Docker inside, root, code interpreter, SSH, public IPv4, HTTPS preview URLs, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, GPU, Python, wake on request, spend limits, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Kaggle Notebooks (kaggle)
https://www.kaggle.com/docs/notebooks · dev-env · isolation: container
- On-demand: 
Features: snapshots, fork/clone, persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, code interpreter, HTTPS preview URLs, egress allowlist, open internet, GPU, Python, MCP server, automatic snapshots, extra volumes, shared volumes

## KakaoCloud Virtual Machine (kakao-cloud)
https://www.kakaocloud.com/services/virtual-machine/pricing · hyperscaler · isolation: vm
- c2a on-demand: c2a.large $0.060386/h; c2a.xlarge $0.120772/h; c2a.2xlarge $0.240808/h
- t1i burstable: t1i.nano $0.00405/h; t1i.micro $0.008174/h; t1i.small $0.016275/h
- m2a general purpose on-demand: m2a.large $0.075114/h; m2a.xlarge $0.150229/h; m2a.2xlarge $0.300458/h
- r2a memory optimized on-demand: r2a.large $0.09868/h; r2a.xlarge $0.19736/h; r2a.2xlarge $0.394719/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, GPU, Windows, live fork (no pause), snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Kamatera (kamatera)
https://www.kamatera.com/pricing/ · hyperscaler · isolation: vm
- Availability shared hourly: A-1c-2g-50GB $0.0136986301369863/h; A-2c-4g-50GB $0.0273972602739726/h; A-4c-8g-50GB $0.0547945205479452/h
- Availability shared monthly: A-1c-2g-50GB $10/h; A-2c-4g-50GB $20/h; A-4c-8g-50GB $40/h
- General purpose reserved thread hourly: B-1c-2g-50GB $0.026027397260273973/h; B-2c-4g-50GB $0.0589041095890411/h; B-4c-8g-50GB $0.12054794520547946/h
- General purpose reserved thread monthly: B-1c-2g-50GB $19/h; B-2c-4g-50GB $43/h; B-4c-8g-50GB $88/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, desktop GUI, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, Windows, Python, Node.js, live resize, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Kasm Workspaces (kasm-workspaces)
https://kasm.com/community-edition · dev-env · isolation: container
- Self-hosted Workspaces licence (Starter: $10/user or $20/concurrent session): Starter licence (compute on your own servers) $null/h
- Kasm-managed Workspaces as a Service (minimum 25 users): Kasm hosted SaaS (quote) $null/h
- Kasm Cloud Personal (hosted disposable browser/desktop, flat monthly with hour cap)
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, Docker inside, root, browser, desktop GUI, preinstalled agents, SSH, HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, arm64, Windows, Python, GPU desktop, live fork (no pause), EU data residency, snapshots on demand, inbound access rules, extra volumes, shared volumes

## Kedge (kedge)
https://kedge.dev/docs/billing · agent-sandbox · isolation: vm
- Persistent machine/workspace (preview price estimates): $0.02054794520547945/vCPU-h, $0.00684931506849315/GiB-h
- Disposable and held sandbox API (public preview): $0.02054794520547945/vCPU-h, $0.00684931506849315/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, private networking, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, webhooks, scheduled runs (cron), EU data residency, scoped API keys, spend limits, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Kernel (kernel)
https://www.onkernel.com/pricing · browser · isolation: firecracker
- Headless: Headless $0.06000012/h
- Headful (8 or 16 GiB): Headful 8 GiB $0.48000096/h; Headful 16 GiB $0.96000192/h
- Headful GPU: GPU $2.88000576/h
- App invocation (serverless agent code, 4 GB): $0/vCPU-h, $0.06000012/GiB-h
Features: snapshots, memory snapshots, pause/resume, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), root, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, SSH, egress allowlist, open internet, static egress IP, BYOC, open source, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, audit logs, EU data residency, zero data retention, spend limits, MCP server, automatic snapshots, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Koyeb Sandboxes (koyeb)
https://www.koyeb.com/pricing · agent-sandbox · isolation: vm
- Eco instances (2 GB/vCPU, older Milan CPUs, was/fra/sin): eco-nano $0.00216/h; eco-micro $0.0036/h; eco-small $0.0072/h
- Standard instances (1 GB/vCPU up to 2xlarge, Genoa, all regions): nano $0.0036/h; micro $0.0072/h; small $0.0144/h
- GPU instances (bundled vCPU/RAM/disk per GPU): $0/vCPU-h, $0/GiB-h
- Savings plan / long-term reservation (up to 50% off, best case): eco-xlarge $0.0576/h; eco-2xlarge $0.1152/h; xlarge $0.1152/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, preinstalled agents, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, private networking, self-hosting, BYOC, SOC 2, SSO, GPU, Python, Node.js, Go, Rust, Java, wake on request, memory fork, audit logs, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## KubeSail (kubesail)
https://kubesail.com · paas · isolation: not published · discontinued

Features: ≥24 h sessions, self-hosting, open source, inbound access rules, firewall inside (nftables)

## Lambda (lambda)
https://lambda.ai/pricing · gpu-cloud · isolation: vm
- B200 SXM6, on-demand (×1): $0/vCPU-h, $0/GiB-h
- GH200, on-demand (×1): $0/vCPU-h, $0/GiB-h
- H100 SXM, on-demand (×1): $0/vCPU-h, $0/GiB-h
- H100 PCIe, on-demand (×1): $0/vCPU-h, $0/GiB-h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, raw TCP inbound, open internet, private networking, SOC 2, SSO, GPU, arm64, Python, Node.js, audit logs, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## LangSmith Sandboxes (langsmith-sandbox)
https://www.langchain.com/pricing · agent-sandbox · isolation: vm
- On-demand: $0.0576/vCPU-h, $0.01845/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, SSH, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, SOC 2, HIPAA, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live fork (no pause), memory fork, audit logs, EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, shared volumes

## Lapdev (lapdev)
https://lap.dev/pricing/ · dev-env · isolation: container
- Native product meter; not a general-purpose 4-vCPU/8-GiB VM: 
Features: fork/clone, pause/resume, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), HTTPS preview URLs, open internet, private networking, BYOC, open source, live fork (no pause), EU data residency, inbound access rules, shared volumes

## Latitude.sh (latitude-sh)
https://www.latitude.sh/pricing · hyperscaler · isolation: dedicated-host
- Bare metal hourly starting rates: c3.small.x86 $0.52/h; m4.metal.small $0.81/h; f4.metal.small $1.09/h
- Virtual machines: vm.small $0.19/h; vm.medium $0.35/h; vm.large $0.65/h
- Bare metal monthly prepaid (measured toggle): c3.small.x86 $190/h; m4.metal.small $296/h; f4.metal.small $398/h
- VM monthly prepaid (measured toggle): vm.small $69/h; vm.medium $128/h; vm.large $237/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, open internet, static egress IP, private networking, SOC 2, SSO, GPU, Windows, live fork (no pause), audit logs, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## Leap0 (leap0)
https://leap0.dev/ · agent-sandbox · isolation: firecracker
- Published rate (not yet charged: free public preview): $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, systemd, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, SSH, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Lightning AI (lightning-ai)
https://lightning.ai/pricing · dev-env · isolation: vm
- The one free Default CPU Studio (4 cores / 16 GB)
- On-demand CPU Studio (fixed shapes): Large (CPU) $0.51/h; X-Large (CPU) $0.99/h; Data prep $1.48/h
- Interruptible CPU Studio: Default (CPU) interruptible $0.22/h; Large (CPU) interruptible $0.33/h; X-Large (CPU) interruptible $0.54/h
- GPU Studio, on-demand (1-GPU rates): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, code interpreter, SSH, HTTPS preview URLs, egress allowlist, private networking, self-hosting, BYOC, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, live fork (no pause), audit logs, EU data residency, spend limits, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Lightpanda Cloud (lightpanda)
https://lightpanda.io/pricing · browser · isolation: container
- Cloud headless browser: Cloud headless browser $0.08/h
Features: idle auto-stop, browser, browser automation API, open internet, self-hosting, BYOC, open source, arm64, Python, Node.js, MCP server

## Limrun (limrun)
https://lim.run · macos · isolation: apple-vm
- Native product; see regime notes
Features: snapshots, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), browser, computer-use API, browser + desktop control, HTTPS preview URLs, static egress IP, SSO, arm64, macOS, Python, Node.js, webhooks, EU data residency, scoped API keys, MCP server, automatic snapshots, gVisor or VM (no shared kernel), inbound access rules

## Lizard (lizard)
https://lizard.build/pricing · agent-sandbox · isolation: firecracker
- Sandbox measured usage (4 vCPU / 4 GiB fixed template): $0.0250128/vCPU-h, $0.0125064/GiB-h
Features: pause/resume, persistent disk, volumes, ≥24 h sessions, full VM (own kernel), root, code interpreter, preinstalled agents, SSH, HTTPS preview URLs, open internet, SOC 2, HIPAA, SSO, Python, Node.js, audit logs, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes

## machine0 (machine0)
https://machine0.io/ · agent-sandbox · isolation: vm
- On-demand: small $0.013/h; medium $0.034/h; large $0.052/h
- GPU VMs, bundled host: gpu-4000ada-1 $0.836/h; gpu-l40s-1 $1.727/h; gpu-6000ada-1 $1.727/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, preinstalled agents, SSH, public IPv4, HTTPS preview URLs, open internet, GPU, Python, Node.js, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## MacinCloud (macincloud)
https://www.macincloud.com/pages/dedicated.html · macos · isolation: vm
- Dedicated Server, Apple silicon (dedicated macOS instance, full admin), monthly: M1, 8 GB, 250 GB SSD (Sonoma) $0.121918/h; M2, 8 GB, 250 GB SSD + RDP add-on $0.156164/h; M4, 16 GB, 250 GB SSD + RDP add-on $0.224658/h
- Pay-as-You-Go by the day (shared managed Mac, no admin): PAYG daily: M1/M2/M4 Mac mini, 8-16 GB (non-admin user account) $0.166667/h
- Managed Server plan (shared managed Mac account), monthly: Managed M1 Mac mini account, 16 GB (starting price) $0.034247/h; Managed M2 Mac mini account, 16 GB (starting price) $0.035616/h; Managed M4 Mac mini account, 16 GB (starting price) $0.039726/h
- Dedicated Server, quarterly/yearly prepaid (effective monthly): Intel 4-core, 8 GB, 120 GB SSD (yearly, effective month) $0.067352/h; M1, 8 GB, 250 GB SSD (yearly, effective month) $0.101598/h; M2, 8 GB, 250 GB SSD + RDP add-on (yearly, effective month) $0.130137/h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, static egress IP, GPU, arm64, macOS, Python, Node.js, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes

## MacStadium (macstadium)
https://www.macstadium.com/pricing · macos · isolation: dedicated-host
- Individual hosted Mac, monthly subscription: M2.S Mac mini: M2 8-core, 8 GB, 256 GB ($109/mo) $0.149315/h; M4.S Mac mini: M4 10-core, 16 GB, 256 GB ($149/mo) $0.20411/h; M2.M Mac mini: M2 8-core, 16 GB, 1 TB ($199/mo) $0.272603/h
- Orka macOS VM cluster (Mac Cloud Compute), contact sales: Orka cluster node (price on request) $null/h
- Orka Burst elastic nodes (add-on to an Orka cluster): Burst node (price on request) $null/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, self-hosting, BYOC, SOC 2, SSO, GPU, arm64, macOS, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Magnitude (magnitude)
https://magnitude.dev/ · browser · isolation: not published · discontinued

Features: ≥24 h sessions, browser automation API, self-hosting, open source

## Manufact (manufact)
https://manufact.com · paas · isolation: firecracker
- Native product; see regime notes
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), HTTPS preview URLs, custom domains, open internet, BYOC, SOC 2, SSO, Python, Node.js, wake on request, EU data residency, spend limits, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules

## Manus Cloud Computer (manus-cloud-computer)
https://help.manus.im/en/articles/15392078-understanding-cloud-computer-plans-and-billing · agent-sandbox · isolation: vm
- Standard server: Standard server $0.041096/h
- Advanced server: Advanced server $0.068493/h
Features: persistent disk, ≥24 h sessions, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, open internet, static egress IP, SOC 2, SSO, Python, Node.js, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables)

## Maritime (maritime)
https://maritime.sh/pricing · agent-sandbox · isolation: firecracker
- Included agent machine: Included agent machine $0/h
- Computer desktop early access: Computer desktop early access $0/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, browser, desktop GUI, computer-use API, desktop control API, browser + desktop control, preinstalled agents, HTTPS preview URLs, Python, Node.js, wake on request, webhooks, scheduled runs (cron), scoped API keys, MCP server, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy

## Mastra (mastra)
https://mastra.ai · paas · isolation: dedicated-vm
- Native product; see regime notes
- Mastra Platform Starter CPU-time meter: CPU hour (hardware entitlement undisclosed) $0.35/h
- Mastra Platform Teams CPU-time meter: CPU hour (hardware entitlement undisclosed) $0.25/h
- Persistent Server 24/7 uptime add-on: per project $null/h
Features: snapshots, fork/clone, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, SSO, Python, Node.js, wake on request, live fork (no pause), audit logs, EU data residency, automatic snapshots, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules

## Metorial (metorial)
https://metorial.com · paas · isolation: container
- Native service
Features: ≥24 h sessions, HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, SSO, Node.js, EU data residency, MCP server, inbound access rules

## microsandbox (microsandbox)
https://microsandbox.dev/pricing · agent-sandbox · isolation: vm
- Cloud microVM (private beta), allocated resources: $0.05/vCPU-h, $0.0162/GiB-h
Features: snapshots, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, systemd, browser, SSH, egress allowlist, open internet, self-hosting, BYOC, open source, SSO, Python, Node.js, credential injection, audit logs, spend limits, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Microsoft Azure GPU VMs (azure-gpu)
https://prices.azure.com/api/retail/prices · gpu-cloud · isolation: vm
- Standard_ND96isr_H100_v5, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- Standard_ND96isr_H100_v5, reserved (×8 node): $0/vCPU-h, $0/GiB-h
- Standard_ND96isr_H100_v5, reserved (×8 node): $0/vCPU-h, $0/GiB-h
- Standard_ND96isr_H100_v5, reserved (×8 node): $0/vCPU-h, $0/GiB-h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Windows, Python, HTTP method/path egress rules, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Microsoft Dev Box (closed to new customers; retiring 2028-09-18) (azure-dev-box)
https://azure.microsoft.com/en-us/products/dev-box/ · windows · isolation: vm · discontinued

Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, Docker inside, nested virtualization, root, browser, desktop GUI, egress allowlist, open internet, private networking, SOC 2, HIPAA, SSO, Windows, audit logs, EU data residency, MCP server, automatic snapshots, snapshots on demand, inbound access rules, firewall inside (nftables)

## Microsoft Foundry hosted agents (East US) (azure-foundry-agents)
https://azure.microsoft.com/en-us/pricing/details/foundry-agent-service/ · hyperscaler · isolation: vm
- East US hosted agent compute: half $0.0615/h; standard $0.123/h; large $0.246/h
- EU hosted agent compute (Sweden Central, France Central, UK South): half $0.06765/h; standard $0.1353/h; large $0.2706/h
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), code interpreter, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, SSO, Python, wake on request, EU data residency, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules

## Minicor (minicor)
https://minicor.com · windows · isolation: dedicated-vm
- Native product; see regime notes
Features: persistent disk, ≥24 h sessions, full VM (own kernel), browser, desktop GUI, computer-use API, browser + desktop control, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, SOC 2, HIPAA, Windows, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## MiniMax Agent hosting / developer API scope (minimax)
https://platform.minimax.io/docs/llms.txt · agent-sandbox · isolation: not published
- Hosted agent environment quote/spec unverified
Features: ≥24 h sessions, Python, Node.js

## MIOSA (miosa)
https://miosa.ai/pricing · agent-sandbox · isolation: firecracker
- Sandbox (fixed sizes xs-xl, managed): xs $0.080667/h; small $0.155187/h; medium $0.310374/h
- Computer (durable Xfce desktop, fixed sizes): xs $0.080667/h; small $0.161334/h; medium $0.328815/h
- Bring your own cloud (45% orchestration fee only, + your cloud bill): xs $0.0363/h; small $0.069834/h; medium $0.139668/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, systemd, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, BYOC, Python, Node.js, credential injection, HTTP method/path egress rules, live fork (no pause), memory fork, webhooks, audit logs, MCP server, automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes

## Mistral Compute / AI Cloud (mistral-compute)
https://mistral.ai/products/aicloud/ · gpu-cloud · isolation: bare-metal
- Dedicated AI capacity
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, SSH, static egress IP, SOC 2, HIPAA, SSO, GPU, arm64, Python, audit logs, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), shared volumes

## Modal (modal)
https://modal.com/pricing · agent-sandbox · isolation: gvisor
- Sandbox (gVisor), request = size, any region: $0.141912/vCPU-h, $0.024012/GiB-h
- Sandbox, minimal request + burst (billed ~ on usage; burst capacity not guaranteed): $0.141912/vCPU-h, $0.024012/GiB-h
- Sandbox pinned to broad region (us / eu / ap): $0.141912/vCPU-h, $0.024012/GiB-h
- Sandbox pinned to narrow region (us-west, uk, jp ...): $0.141912/vCPU-h, $0.024012/GiB-h
Features: snapshots, memory snapshots, fork/clone, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, nested virtualization, root, systemd, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, SOC 2, HIPAA, SSO, GPU, Python, Node.js, memory fork, audit logs, EU data residency, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Modelence (modelence)
https://modelence.com · paas · isolation: container
- App containers: Micro $0.013/h; Micro Plus $0.02/h; Small $0.03/h
Features: pause/resume, ≥24 h sessions, HTTPS preview URLs, open source, Node.js, live resize, EU data residency, MCP server

## Moonshot Kimi Hosted Agents sandbox (moonshot-kimi)
https://platform.kimi.com/docs/pricing/hosted-agents.md · agent-sandbox · isolation: not published
- Hosted Agents 1C/1G sandbox: 1C1GB $0.010729463537/h
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, browser, code interpreter, preinstalled agents, egress allowlist, open internet, Python, Node.js, Go, Rust, credential injection, wake on request, scheduled runs (cron), agent harness API, hosted agent API (their own agent), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Morph Cloud (morph-cloud)
https://cloud.morph.so/web/subscribe · agent-sandbox · isolation: vm
- MorphVM instance / Devbox, pay-as-you-go MCU ($0.05): $0.05/vCPU-h, $0/GiB-h
- Morph Actions GitHub runner (fixed sizes, billed per runner-hour, no idle pause): Small $0.15/h; Medium $0.25/h; Large $0.5/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, desktop GUI, SSH, HTTPS preview URLs, open internet, Windows, macOS, Python, Node.js, wake on request, live fork (no pause), memory fork, audit logs, scoped API keys, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, extra volumes, shared volumes

## Mosaic Sandbox (mosaic)
https://sandbox.mosaicos.com/ · agent-sandbox · isolation: firecracker
- Active compute, $0.05 per billing unit per awake hour (hibernated = $0): 1 vCPU / 2048 MiB (bills as 1 unit) $0.05/h; 2 vCPU / 4096 MiB (default, 1 unit, kept warm) $0.05/h; 4 vCPU / 8192 MiB (2 units, cold boot) $0.1/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, SSO, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, audit logs, scoped API keys, MCP server, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Namespace (namespace)
https://namespace.so/pricing · dev-env · isolation: vm
- Linux compute instance (unit-minutes, PAYG/overage $0.0015): 1x2 $0.09/h; 2x4 $0.18/h; 2x8 $0.36/h
- Windows compute instance (2x multiplier, Team+ only): win-1x2 $0.18/h; win-2x4 $0.36/h; win-2x8 $0.72/h
- macOS instance on Apple Silicon (10x multiplier): mac-4x7 $3.6/h; mac-6x14 $5.4/h; mac-6x28 $6.75/h
- Linux arm64 on Apple Silicon (7x multiplier, early access): as-6x14 $4.41/h; as-12x28 $8.82/h; as-12x56 $17.64/h
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, desktop GUI, computer-use API, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, SSO, arm64, Windows, macOS, Python, Node.js, credential injection, wake on request, audit logs, EU data residency, spend limits, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes

## NAVER Cloud / LINE-NAVER scope (naver-cloud)
https://www.ncloud.com/product/compute/server · hyperscaler · isolation: vm
- Korea VPC g3 High CPU: High CPU-g3 4/8 $0.141391929064/h
- Korea VPC g2 High CPU: High CPU-g2 4/8 50GB $0.144337594253/h
- Korea VPC g3 High CPU monthly: High CPU-g3 4/8 $null/h
- Korea VPC g2 High CPU 50GB monthly: High CPU-g2 4/8 50GB $null/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, static egress IP, private networking, SOC 2, SSO, GPU, Windows, Python, Node.js, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Nebius (nebius)
https://docs.nebius.com/compute/resources/pricing · gpu-cloud · isolation: vm
- B300, on-demand: $0/vCPU-h, $0/GiB-h
- B300, spot: $0/vCPU-h, $0/GiB-h
- B200, on-demand: $0/vCPU-h, $0/GiB-h
- B200, spot: $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Python, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Nebius ConTree (Token Factory Sandboxes) (nebius-contree)
https://tokenfactory.nebius.com/sandboxes/about · agent-sandbox · isolation: vm
- Sandboxes beta (free, per-command microVM executions)
Features: snapshots, fork/clone, persistent disk, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, egress allowlist, SOC 2, Python, Node.js, EU data residency, scoped API keys, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables)

## netcup VPS (netcup)
https://www.netcup.com/en/server/vps · hyperscaler · isolation: vm
- G12.5 VPS 1-month term (net EUR converted): VPS 500 G12.5 $8.510744/h; VPS 1000 G12.5 $15.371678/h; VPS 2000 G12.5 $29.025278/h
- G12.5 VPS 12-month term (net EUR converted): VPS 500 G12.5 $7.327432/h; VPS 1000 G12.5 $13.289503999999999/h; VPS 2000 G12.5 $25.168136/h
- G12.5 VPS 24-month term (net EUR converted): VPS 500 G12.5 $6.14412/h; VPS 1000 G12.5 $11.207329999999999/h; VPS 2000 G12.5 $21.310993999999997/h
- G12.5 VPS 1-month term, Manassas USA (net EUR converted): VPS 500 G12.5 $9.876104/h; VPS 1000 G12.5 $17.761058/h; VPS 2000 G12.5 $33.462698/h
Features: snapshots, memory snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, GPU, arm64, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Nextmv (nextmv)
https://nextmv.io · paas · isolation: container
- Native product; see regime notes: 6c9500mb870s (standard, Innovator pay-as-you-go) $9.72/h
Features: SSO, GPU, arm64, Python, webhooks, MCP server, inbound access rules

## Nodus Compute (nodus-compute)
https://www.nodus-compute.ai/pricing/ · agent-sandbox · isolation: not published
- Admitted managed sandbox
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, HTTPS preview URLs, egress allowlist, BYOC, GPU, Python, Node.js, wake on request, MCP server, automatic snapshots, hosted agent API (their own agent), inbound access rules, secret proxy, extra volumes

## Northflank Sandboxes (northflank)
https://northflank.com/pricing · agent-sandbox · isolation: vm
- Managed cloud, predefined compute plans: nf-compute-10 $0.0038/h; nf-compute-20 $0.0075/h; nf-compute-50 $0.0167/h
- Managed cloud, custom vCPU/RAM plan (on request): $0.01667/vCPU-h, $0.00833/GiB-h
- BYOC fee only (+ your cloud bill for VMs/GPUs, disk, egress): $0.01389/vCPU-h, $0.00139/GiB-h
- Managed cloud GPU sandbox (gVisor), combined GPU rate: $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, self-hosting, BYOC, SOC 2, HIPAA, SSO, GPU, arm64, Python, Node.js, live fork (no pause), webhooks, scheduled runs (cron), audit logs, EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Notte (notte)
https://www.notte.cc/pricing · browser · isolation: container
- Browser session: Browser session $0/h
- Browser function: Browser function $0.05/h
Features: idle auto-stop, ≥24 h sessions, browser, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, static egress IP, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, Python, EU data residency, MCP server, hosted agent API (their own agent), inbound access rules

## Novita AI Agent Sandbox (novita)
https://docs.novita.ai/guides/sandbox-pricing · agent-sandbox · isolation: firecracker
- Pay-as-you-go (prepaid balance): $0.03528/vCPU-h, $0.01152/GiB-h
- GPU Instance on-demand (H100 SXM 80GB, L40S, RTX 4090, RTX 5090, RTX 6000 Ada): $0/vCPU-h, $0/GiB-h
- GPU Instance spot (H100 SXM 80GB, RTX 4090, RTX 6000 Ada): $0/vCPU-h, $0/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, systemd, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, SOC 2, Python, Node.js, credential injection, wake on request, live resize, memory fork, audit logs, automatic snapshots, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## OakHost (oakhost)
https://www.oakhost.com/mac-mini-hosting · macos · isolation: dedicated-host
- Dedicated Mac mini, monthly (cancel any time): MAC M1.S: M1 8-core, 8 GB, 256 GB (EUR 70/mo) $0.112192/h; MAC M1.M: M1 8-core, 16 GB, 256 GB (EUR 80/mo) $0.128219/h; MAC M1.L: M1 8-core, 16 GB, 1 TB (EUR 85/mo) $0.136233/h
- "Try macOS for a week" (M2, 8+ GB), one-off 7 days: Try macOS: M2, 8+ GB, 256+ GB (EUR 30 per 7 days) $0.208929/h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, raw TCP inbound, egress allowlist, open internet, static egress IP, arm64, macOS, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Okteto (okteto)
https://okteto.com · dev-env · isolation: container
- Native product; see regime notes
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, open source, SOC 2, SSO, GPU, arm64, Python, Node.js, wake on request, EU data residency, snapshots on demand, inbound access rules, extra volumes, shared volumes

## Omnara (omnara)
https://www.omnara.com/pricing · agent-sandbox · isolation: not published
- Managed machines: $0/vCPU-h, $0.0414/GiB-h
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, self-hosting, BYOC, open source, SSO, arm64, wake on request, webhooks, scheduled runs (cron), audit logs, MCP server, automatic snapshots, hosted agent API (their own agent)

## OmniRun (omnirun)
https://omnirun.io/pricing · agent-sandbox · isolation: firecracker
- Launch plan sandbox-hours: Standard sandbox-hour (up to 4 vCPU / 2 GB) $0.2162/h
- Scale plan sandbox-hours: Standard sandbox-hour (up to 4 vCPU / 2 GB) $0.1707/h
Features: snapshots, pause/resume, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, open source, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, webhooks, audit logs, EU data residency, spend limits, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy

## Ona (formerly Gitpod) (ona)
https://ona.com/pricing · dev-env · isolation: vm
- Ona Cloud (Core), Standard 4 vCPU / 16 GB = 1 OCU/h: Standard (4 vCPU / 16 GB, 1 OCU/h) $0.25/h
- Ona Cloud (Core) GPU VM, 16 vCPU / 64 GB = 7 OCU/h: GPU / Data Science VM (16 vCPU / 64 GB / 300 GB, 7 OCU/h) $1.75/h
- Enterprise runner in own AWS VPC (customer pays EC2 on-demand): Small (m6i.large) $0.096/h; Regular (m6i.xlarge) $0.192/h; Large (m6i.2xlarge) $0.384/h
- Enterprise runner in own AWS VPC, GPU Large (g5.4xlarge, 1× A10G): GPU Large (g5.4xlarge, 1× A10G 24 GB) $1.624/h
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, preinstalled agents, SSH, HTTPS preview URLs, open internet, BYOC, SOC 2, SSO, GPU, webhooks, scheduled runs (cron), EU data residency, spend limits, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## OneCLI (onecli)
https://onecli.sh · agent-sandbox · isolation: vm
- Native product; see regime notes
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, browser, preinstalled agents, SSH, egress allowlist, open internet, self-hosting, open source, SSO, arm64, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, scheduled runs (cron), audit logs, automatic snapshots, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy, secret proxy for any API

## Open Telekom Cloud / T Cloud Public (open-telekom-cloud)
https://www.open-telekom-cloud.com/en/prices · hyperscaler · isolation: vm
- ECS s9 Linux on demand: s9.xlarge.2 $0.2148405338/h
- Flexible purpose x1e.xlarge.2 EU-DE: x1e.xlarge.2 $0.1309243704/h
- s9.xlarge.2 reserved 12 months
- s9.xlarge.2 reserved 24 months
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, static egress IP, private networking, SOC 2, SSO, GPU, Windows, Python, Node.js, HTTP method/path egress rules, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## OpenAGI / Lux (openagi)
https://developer.agiopen.org/docs/pricing · browser · isolation: not published
- Bring your own computer: Bring your own computer $null/h
Features: ≥24 h sessions, desktop GUI, computer-use API, desktop control API, self-hosting, BYOC, open source, HIPAA, arm64, Python, Node.js, hosted agent API (their own agent)

## OpenAI Containers (Code Interpreter / Hosted Shell) (openai-containers)
https://developers.openai.com/api/docs/pricing · agent-sandbox · isolation: vm
- Responses API container (Code Interpreter / Hosted Shell), per minute with 5-min minimum: memory_limit 1g $0.09/h; memory_limit 4g $0.36/h; memory_limit 16g $1.44/h
- Agents API OpenAI-hosted sandbox (public beta), small/medium/large at container rates: small $0.09/h; medium (default) $0.36/h; large $1.44/h
- Codex Cloud task VM (bundled in ChatGPT seat; tokens/credits only): Plus / Edu Plus task VM $null/h; Pro / Business / Enterprise / Edu task VM $null/h
- ChatGPT agent mode / Work cloud browser (per message, credits): cloud computer + browser (size unpublished) $null/h
Features: idle auto-stop, ≥24 h sessions, full VM (own kernel), browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, egress allowlist, self-hosting, SOC 2, HIPAA, SSO, Python, Node.js, Go, Java, credential injection, EU data residency, agent harness API, gVisor or VM (no shared kernel), secret proxy, secret proxy for any API

## OpenComputer (opencomputer)
https://opencomputer.dev/sandboxes · agent-sandbox · isolation: vm
- Machine time per second (2 GB/1 vCPU default, auto-burst to 4/2 or 8/4): 0.25 vCPU / 0.5 GB (bursts to 1/2) $0.0474/h; 0.5 vCPU / 1 GB (bursts to 2/4) $0.0948/h; default 1 vCPU / 2 GB (bursts to 4/8) $0.189/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, browser automation API, anti-bot stealth, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, open source, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, webhooks, scheduled runs (cron), audit logs, MCP server, automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, shared volumes

## OpenHands Remote Sandbox / Cloud (openhands-runtime)
https://docs.openhands.dev/openhands/usage/sandboxes/remote · dev-env · isolation: container
- Managed OpenHands Cloud
- Remote agent server on own infrastructure
Features: pause/resume, persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, nested virtualization, browser, preinstalled agents, HTTPS preview URLs, self-hosting, BYOC, open source, SSO, Python, Node.js, wake on request, EU data residency, spend limits, agent harness API, inbound access rules, secret proxy, extra volumes

## OpenPond Cloud Sandboxes (openpond)
https://openpond.ai/pricing · agent-sandbox · isolation: firecracker
- On-demand: $0.072/vCPU-h, $0.0234/GiB-h: default (1 vCPU / 2 GiB / 10 GiB) $0.11988/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, HTTPS preview URLs, egress allowlist, open source, SSO, Node.js, live fork (no pause), scheduled runs (cron), scoped API keys, spend limits, MCP server, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## OpenReward Sandboxes (openreward)
https://openreward.ai/pricing · agent-sandbox · isolation: container
- Pay-as-you-go, per second: $0.0504/vCPU-h, $0.0162/GiB-h
Features: volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, egress allowlist, open internet, self-hosting, GPU, Python, Node.js, credential injection, secret proxy, secret proxy for any API, extra volumes, shared volumes

## Opensteer (opensteer)
https://opensteer.com/pricing · browser · isolation: not published
- Browser / computer runtime (per hour): browser runtime $0.2/h; computer runtime $0.2/h
Features: persistent disk, ≥24 h sessions, browser, desktop GUI, computer-use API, browser automation API, browser + desktop control, preinstalled agents, self-hosting, open source, Python, scheduled runs (cron), agent harness API, hosted agent API (their own agent)

## Oracle Cloud Infrastructure (oracle-cloud)
https://www.oracle.com/cloud/compute/pricing/ · hyperscaler · isolation: vm
- E4 Flex x86: $0.0125/vCPU-h, $0.0015/GiB-h: 
- E5/E6 Flex x86: $0.015/vCPU-h, $0.002/GiB-h: 
- Intel Standard3 Flex: $0.02/vCPU-h, $0.0015/GiB-h: 
- Ampere A1 Arm: $0.01/vCPU-h, $0.0015/GiB-h: 
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, Windows, macOS, HTTP method/path egress rules, EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Orgo (orgo)
https://www.orgo.ai/pricing · agent-sandbox · isolation: vm
- Monthly plan pool (Linux)
- Monthly plan pool (Windows, Scale plan only)
- Yearly billing (Linux)
- Yearly billing (Windows, Scale only)
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, residential IPs, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, static egress IP, self-hosting, BYOC, GPU, Windows, macOS, Python, Node.js, GPU desktop, credential injection, wake on request, live resize, live fork (no pause), memory fork, webhooks, scoped API keys, spend limits, MCP server, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## orkestr Sandboxes (orkestr)
https://orkestr.eu/sandboxes · agent-sandbox · isolation: vm
- Pay-as-you-go sandbox (metered on actual use): $0.051201/vCPU-h, $0.017067/GiB-h
Features: snapshots, memory snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, HTTPS preview URLs, egress allowlist, open internet, SSO, Python, Node.js, wake on request, EU data residency, MCP server, automatic snapshots, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## OVHcloud GPU instances (ovh-gpu)
https://www.ovhcloud.com/en/public-cloud/prices/ · gpu-cloud · isolation: vm
- l40s-90, on-demand (×1): $0/vCPU-h, $0/GiB-h
- a10-45, on-demand (×1): $0/vCPU-h, $0/GiB-h
- a100-180, on-demand (×1): $0/vCPU-h, $0/GiB-h
- h100-380, on-demand (×1): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, SOC 2, HIPAA, SSO, GPU, Windows, live fork (no pause), EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## OVHcloud Public Cloud (ovhcloud)
https://us.ovhcloud.com/public-cloud/prices/ · hyperscaler · isolation: vm
- B3 general purpose hourly: b3-8 $0.069105/h; b3-16 $0.13801/h; b3-32 $0.27592/h
- C3 compute optimized hourly: c3-4 $0.062605/h; c3-8 $0.12501/h; c3-16 $0.24982/h
- Gen2 monthly subscription: b2-15 $59.54/h; c2-15 $87.89/h
- Gen3 12-month Savings Plan: b3-8 $0.06003/h; b3-16 $0.11989/h; b3-32 $0.239695/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Windows, Python, live fork (no pause), EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## PandaStack (pandastack)
https://www.pandastack.ai/pricing/ · agent-sandbox · isolation: firecracker
- Hosted8-vCPU burstable template; active CPU and resident RAM: $0.054/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, code interpreter, preinstalled agents, HTTPS preview URLs, open internet, self-hosting, BYOC, open source, SSO, Python, Node.js, Go, wake on request, live fork (no pause), memory fork, webhooks, spend limits, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## PaperPod (paperpod)
https://www.paperpod.dev/ · agent-sandbox · isolation: container
- Native product meter; not a general-purpose 4-vCPU/8-GiB VM: 
Features: idle auto-stop, ≥24 h sessions, browser, code interpreter, HTTPS preview URLs, open internet, Python, Node.js, wake on request

## Paperspace (paperspace)
https://docs.digitalocean.com/products/paperspace/pricing/ · gpu-cloud · isolation: vm
- Current Core CPU hourly: C1 $0.0045/h; C3 $0.018/h; C4 $0.04/h
- RTX-A4000, on-demand (×1): $0/vCPU-h, $0/GiB-h
- RTX-A5000, on-demand (×1): $0/vCPU-h, $0/GiB-h
- RTX-A6000, on-demand (×1): $0/vCPU-h, $0/GiB-h
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Python, EU data residency, spend limits, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## Party (party)
https://party.build · agent-sandbox · isolation: container
- Native product; see regime notes
Features: ≥24 h sessions, preinstalled agents, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, HIPAA, SSO, Python, Node.js, EU data residency, spend limits, inbound access rules

## Pipekit (pipekit)
https://pipekit.io/pricing · dev-env · isolation: container
- Native product; see regime notes
Features: ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, self-hosting, BYOC, SSO, arm64, Python, scheduled runs (cron), audit logs, EU data residency, MCP server

## Pipeshift (pipeshift)
https://pipeshift.com · gpu-cloud · isolation: dedicated-vm
- Native product; see regime notes
Features: ≥24 h sessions, full VM (own kernel), BYOC, SOC 2, GPU, EU data residency, gVisor or VM (no shared kernel)

## Playgent (playgent)
https://useplaygent.com · agent-sandbox · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, MCP server

## Ploomber (ploomber)
https://ploomber.io/ · paas · isolation: not published
- Native product; see regime notes
Features: idle auto-stop, ≥24 h sessions, HTTPS preview URLs, SSO, Python, inbound access rules

## PoplarML (poplarml)
http://poplarml.com · gpu-cloud · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, HTTPS preview URLs, GPU, Python

## Porter (porter)
https://porter.run · paas · isolation: container
- Native service
- Porter Sandboxes (AWS, in your own account)
Features: snapshots, fork/clone, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, HTTPS preview URLs, custom domains, egress allowlist, static egress IP, self-hosting, BYOC, SOC 2, HIPAA, SSO, GPU, Python, Node.js, live fork (no pause), scheduled runs (cron), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, extra volumes, shared volumes

## PPIO Agent Sandbox (ppio-sandbox)
https://ppio.com/docs/sandbox/pricing.md · agent-sandbox · isolation: firecracker
- CPU + RAM allocation: $0.016094195306/vCPU-h, $0.008047097653/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, static egress IP, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Prime Intellect Sandboxes (prime-intellect)
https://www.primeintellect.ai/sandboxes · agent-sandbox · isolation: vm
- VM sandbox, launch pricing (valid through 2026-12-22): $0.02/vCPU-h, $0.0125/GiB-h
- GPU Pod marketplace, H100 80GB PCIe on-demand (cheapest live offer 2026-09-29: Lambda, 26 vCPU / 200 GB): $0/vCPU-h, $0/GiB-h
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, systemd, SSH, egress allowlist, open internet, SOC 2, GPU, EU data residency, spend limits, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Prized (prized)
https://prized.dev/docs/billing · dev-env · isolation: vm
- On-demand: Nano $0.0136986301369863/h; Micro $0.03424657534246575/h; Extra Small $0.0684931506849315/h
Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, browser, desktop GUI, preinstalled agents, SSH, public IPv4, HTTPS preview URLs, open internet, Python, wake on request, live fork (no pause), audit logs, spend limits, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Railway (railway)
https://railway.com/pricing · agent-sandbox · isolation: vm
- Sandboxes (VM rates): $0.069444/vCPU-h, $0.069444/GiB-h
- Services (container rates, always-on or serverless service): $0.027792/vCPU-h, $0.013896/GiB-h
- Cloud Agents (VM rates, fixed size, beta): $0.069444/vCPU-h, $0.069444/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, browser, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, open internet, static egress IP, private networking, BYOC, SOC 2, HIPAA, SSO, Python, Node.js, live fork (no pause), webhooks, scheduled runs (cron), audit logs, EU data residency, spend limits, MCP server, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Red Hat OpenShift (Harbor backend) (openshift)
https://www.redhat.com/en/technologies/cloud-computing/openshift/pricing · paas · isolation: container
- OpenShift cluster pods
- OpenShift Dedicated worker subscription component: $0.04275/vCPU-h, $0/GiB-h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, HTTPS preview URLs, egress allowlist, static egress IP, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, arm64, Windows, Python, Node.js, live resize, EU data residency, snapshots on demand, inbound access rules, extra volumes, shared volumes

## Reflex (reflex)
https://reflex.dev/pricing/ · paas · isolation: container
- Hosted Python apps
Features: fork/clone, idle auto-stop, ≥24 h sessions, HTTPS preview URLs, custom domains, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, Python, Node.js, wake on request, audit logs, EU data residency, MCP server

## Refresh (refresh)
https://www.refresh.dev · agent-sandbox · isolation: not published
- Native service
Features: ≥24 h sessions, browser, desktop GUI, computer-use API, browser + desktop control, self-hosting, open source

## Release (release)
https://release.com/pricing · dev-env · isolation: container
- Enterprise BYOC
Features: snapshots, memory snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, HTTPS preview URLs, custom domains, open internet, static egress IP, BYOC, SOC 2, SSO, GPU, arm64, Windows, Python, Node.js, scheduled runs (cron), EU data residency, snapshots on demand, inbound access rules, extra volumes, shared volumes

## Remote Browser (remote-browser)
https://remote-browser.dev/pricing · browser · isolation: container
- Purchased browser-hour packs: Purchased browser-hour packs $0.245/h
Features: idle auto-stop, ≥24 h sessions, browser, computer-use API, browser automation API, anti-bot stealth, residential IPs, preinstalled agents, open internet, EU data residency, scoped API keys

## Render (render)
https://render.com/pricing · paas · isolation: container
- Web / private service or background worker (fixed compute plan, prorated per second): 0.5c-512mb (Starter) $0.009589/h; 1c-2g (Standard) $0.034247/h; 2c-4g (Pro) $0.116438/h
- Cron job (billed per second while running): 0.5c-512mb $0.0096/h; 1c-2g $0.0348/h; 2c-4g $0.1182/h
- Workflow task run, fixed compute plan (per second while running): 2c-4g $0.4/h; 2c-8g $0.7/h; 4c-8g $1/h
- Workflow task run, flex plan (billed on actual CPU time + sampled RAM): $0.2/vCPU-h, $0.05/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, Python, Node.js, Go, Rust, live fork (no pause), memory fork, webhooks, scheduled runs (cron), audit logs, EU data residency, MCP server, snapshots on demand, inbound access rules, extra volumes

## RentaMac (rentamac.io) (rentamac)
https://rentamac.io/pricing · macos · isolation: dedicated-host
- Daily rental ($19/day): Dedicated Mac mini M4 10-core, 16 GB, 256 GB (daily) $0.791667/h
- Weekly rental ($59/week): Dedicated Mac mini M4 10-core, 16 GB, 256 GB (weekly) $0.35119/h
- Monthly rental ($119/month): Dedicated Mac mini M4 10-core, 16 GB, 256 GB (monthly) $0.163014/h
- 3-month prepay ($339 / 3 months): Dedicated Mac mini M4 10-core, 16 GB, 256 GB (3-month prepay) $0.154795/h
Features: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, open internet, arm64, macOS, Python, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes

## Replicas (replicas)
https://replicas.dev · agent-sandbox · isolation: vm
- Team Flex manual workspaces: Team $0.96/h
- Developer Flex manual workspace: small $0.48/h
- API and automation workspace: small $0.48/h; large $0.96/h
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, open internet, static egress IP, self-hosting, SOC 2, Python, Node.js, wake on request, scheduled runs (cron), audit logs, MCP server, automatic snapshots, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Replicate (replicate)
https://replicate.com/pricing · gpu-cloud · isolation: container
- Private Cog CPU model: cpu-small $0.09/h; cpu $0.36/h
- T4, serverless (×1): $0/vCPU-h, $0/GiB-h
- A100-80GB, serverless (×1): $0/vCPU-h, $0/GiB-h
- H100, serverless (×1): $0/vCPU-h, $0/GiB-h
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, HTTPS preview URLs, self-hosting, open source, GPU, Python, Node.js, wake on request, webhooks, audit logs, spend limits, MCP server

## Replit (replit)
https://docs.replit.com/billing/deployment-pricing · paas · isolation: vm
- Reserved VM deployment (always-on shared/dedicated VM, hourly): shared 0.5 vCPU / 2 GiB $0.0208/h; dedicated 1 vCPU / 4 GiB $0.0486/h; dedicated 2 vCPU / 8 GiB $0.0694/h
- Autoscale / Scheduled deployment (compute units while serving or running): $0.03888/vCPU-h, $0.00432/GiB-h
Features: snapshots, fork/clone, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), desktop GUI, preinstalled agents, SSH, HTTPS preview URLs, custom domains, open internet, static egress IP, SOC 2, SSO, Python, Node.js, Go, Java, wake on request, scheduled runs (cron), audit logs, EU data residency, spend limits, MCP server, automatic snapshots, agent harness API, gVisor or VM (no shared kernel), inbound access rules, secret proxy

## Rescale (rescale)
https://rescale.com · gpu-cloud · isolation: dedicated-vm
- Native product; see regime notes
- On-Demand Priority
- On-Demand Economy
- On-Demand Reserved
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, desktop GUI, SSH, public IPv4, HTTPS preview URLs, egress allowlist, BYOC, SOC 2, HIPAA, SSO, GPU, arm64, Windows, Python, EU data residency, spend limits, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Revyl (revyl)
https://www.revyl.com · macos · isolation: container
- Native product; see regime notes
Features: persistent disk, idle auto-stop, ≥24 h sessions, computer-use API, SSH, static egress IP, SOC 2, HIPAA, SSO, arm64, macOS, Python, Node.js, audit logs, MCP server, hosted agent API (their own agent)

## Rivet (Actors & agentOS) (rivet)
https://rivet.dev/pricing/ · agent-sandbox · isolation: v8-isolate + wasm (agentOS); processes (Actors)
- agentOS sandbox (V8 isolates + WebAssembly): $0.1188/vCPU-h, $0.01044/GiB-h
- Rivet Actors (Node / Bun processes): $0.1188/vCPU-h, $0.01044/GiB-h
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, open source, Python, Node.js, wake on request, MCP server, agent harness API, gVisor or VM (no shared kernel), inbound access rules, extra volumes

## Riza Code Interpreter (riza)
https://riza.io/pricing · agent-sandbox · isolation: isolate · discontinued

Features: egress allowlist, open internet, self-hosting, BYOC, SOC 2, Python, credential injection, MCP server, secret proxy, secret proxy for any API

## Run Cloud (run-cloud)
https://docs.run.cloud/sandboxes/index.md · agent-sandbox · isolation: firecracker · discontinued

Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, desktop GUI, computer-use API, SSH, HTTPS preview URLs, custom domains, spend limits, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## RunAnywhere (runanywhere)
https://www.runanywhere.ai/ · inference-api · isolation: not published
- Wally hosted token-billed inference preview (not an arbitrary-code VM)
Features: ≥24 h sessions, preinstalled agents, self-hosting, open source, SOC 2, Python, Node.js, zero data retention

## RunKit (runkit)
https://runkit.com/ · agent-sandbox · isolation: not published · discontinued

Features: ≥24 h sessions

## Runloop (runloop)
https://www.runloop.ai/pricing · agent-sandbox · isolation: vm
- Devbox CUSTOM_SIZE (allocated, per-second): $0.108/vCPU-h, $0.0252/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, private networking, BYOC, SOC 2, HIPAA, SSO, arm64, Python, Node.js, credential injection, wake on request, live fork (no pause), scoped API keys, MCP server, automatic snapshots, snapshots on demand, agent harness API, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Runpod (runpod)
https://www.runpod.io/pricing · gpu-cloud · isolation: container
- CPU Pod, Secure Cloud, compute-optimized (cpu3c): cpu3c-2-4 $0.06/h; cpu3c-4-8 $0.12/h; cpu3c-8-16 $0.24/h
- GPU Pod, Secure Cloud on-demand (per GPU, vCPU/RAM bundled): $0/vCPU-h, $0/GiB-h
- GPU Pod, Community Cloud on-demand (peer-to-peer hosts, variable reliability): $0/vCPU-h, $0/GiB-h
- Serverless flex worker (GPU, scale to zero): $0/vCPU-h, $0/GiB-h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, root, code interpreter, SSH, public IPv4, HTTPS preview URLs, raw TCP inbound, open internet, private networking, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, webhooks, audit logs, EU data residency, scoped API keys, spend limits, MCP server, inbound access rules, extra volumes, shared volumes

## RunsOn (runs-on)
https://runs-on.com/pricing/ · paas · isolation: vm
- Spot EC2 runner, x64: 1cpu t8i.medium $0.036/h; 2cpu m7i-flex.large $0.054/h; 4cpu m7i-flex.xlarge $0.084/h
- On-demand EC2 runner, x64: 1cpu t8i.medium $0.066/h; 2cpu m7i-flex.large $0.114/h; 4cpu m7i-flex.xlarge $0.21/h
- Spot EC2 runner, arm64 (Graviton): 1cpu m8g.medium $0.036/h; 2cpu m8g.large $0.06/h; 4cpu m8g.xlarge $0.096/h
- On-demand EC2 runner, arm64 (Graviton): 1cpu m8g.medium $0.06/h; 2cpu m8g.large $0.108/h; 4cpu m8g.xlarge $0.198/h
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, SSH, public IPv4, egress allowlist, static egress IP, private networking, self-hosting, BYOC, GPU, arm64, Windows, Python, Node.js, wake on request, EU data residency, automatic snapshots, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Runta (runta)
https://runta.com/pricing/ · agent-sandbox · isolation: vm
- On-demand: $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, scoped API keys, automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Runtime (withruntime.com) (withruntime)
https://withruntime.com/pricing · agent-sandbox · isolation: firecracker
- Active CPU plus reserved RAM: $0.025/vCPU-h, $0.0075/GiB-h
- Reserved CPU: $0.025/vCPU-h, $0.0075/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, SSH, public IPv4, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SSO, Python, Node.js, wake on request, memory fork, webhooks, audit logs, scoped API keys, spend limits, MCP server, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Sail Research (Sailboxes) (sail)
https://docs.sailresearch.com/pricing · agent-sandbox · isolation: firecracker
- Sailbox, billed on observed usage (used vCPU / used RAM / used disk): $0.015/vCPU-h, $0.008/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, systemd, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, egress allowlist, open internet, SOC 2, HIPAA, arm64, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live fork (no pause), memory fork, scheduled runs (cron), zero data retention, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Sakura Internet Cloud (sakura-cloud)
https://cloud.sakura.ad.jp/products/server/ · hyperscaler · isolation: vm
- Normal VM, displayed Ishikari zone: 1C1GB $0.040563/h; 1C2GB $0.052153/h; 2C2GB $0.092716/h
- Normal VM + 12-month discount passport (Ishikari 1): 4C8GB $null/h
- Core-dedicated plan (Intel Xeon), all zones same price: 2C4GB $0.289738/h; 4C8GB $0.58527/h; 8C32GB $1.477662/h
- 高火力 VRT GPU VM (Ishikari 1 only): H100x1 $5.736807/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, static egress IP, SOC 2, SSO, GPU, Windows, Python, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Samsung SDS Cloud Platform (samsung-cloud)
https://cloud.samsungsds.com/serviceportal/pricing.html · hyperscaler · isolation: vm
- Virtual Server, no contract (hourly, open-source OS): s1v1m2 (1st Gen) $0.0443/h; s1v2m4 (1st Gen) $0.1026/h; s1v2m8 (1st Gen) $0.107/h
- Planned Compute 1 year: s1v1m2 (1st Gen) $0.02871/h; s1v2m4 (1st Gen) $0.06659/h; s1v2m8 (1st Gen) $0.06933/h
- Planned Compute 3 years: s1v1m2 (1st Gen) $0.02644/h; s1v2m4 (1st Gen) $0.06147/h; s1v2m8 (1st Gen) $0.06408/h
- Virtual Server, no contract, Windows: s1v1m2 (1st Gen) $0.0565/h; s1v2m4 (1st Gen) $0.1261/h; s1v2m8 (1st Gen) $0.1304/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SSO, GPU, Windows, live resize, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Sandbox as a Service (sandbox-as-a-service)
https://sandbox-as-a-service.com/pricing · agent-sandbox · isolation: vm
- On-demand: small $0.09/h; medium $0.28/h; large $0.55/h
Features: snapshots, fork/clone, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), HTTPS preview URLs, open internet, Python, Node.js, live fork (no pause), EU data residency, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules

## Sandbox0 (sandbox0)
https://sandbox0.ai/pricing · agent-sandbox · isolation: gvisor
- Sandbox0 Cloud, memory-only meter (CPU included): $0/vCPU-h, $0.015/GiB-h
- Managed Agents (Sandpi Platform): sandbox runtime free, pay model tokens + 5% top-up fee: $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, Docker inside, root, browser, desktop GUI, code interpreter, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, self-hosting, open source, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, webhooks, audit logs, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, secret proxy, secret proxy for any API

## Scaleway Apple silicon (Mac mini) (scaleway-apple-silicon)
https://www.scaleway.com/en/pricing/apple-silicon/ · macos · isolation: dedicated-host
- Hourly, no commitment (24 h minimum for macOS): M1-M: M1 8-core, 8 GB, 256 GB (EUR 0.11/h) $0.1287/h; M2-M: M2 8-core, 16 GB, 256 GB (EUR 0.17/h) $0.1989/h; M2-L: M2 Pro 10-core, 16 GB, 512 GB (EUR 0.21/h) $0.2457/h
- Monthly commitment (auto-renewing): M1-M: M1 8-core, 8 GB, 256 GB (EUR 75/month) $0.120205/h; M2-M: M2 8-core, 16 GB, 256 GB (EUR 115/month) $0.184315/h; M2-L: M2 Pro 10-core, 16 GB, 512 GB (EUR 139/month) $0.222781/h
Features: persistent disk, ≥24 h sessions, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, SSH, public IPv4, raw TCP inbound, static egress IP, private networking, SSO, arm64, macOS, Python, Node.js, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables)

## Scaleway GPU instances (scaleway-gpu)
https://www.scaleway.com/en/pricing/gpu/ · gpu-cloud · isolation: vm
- B300-SXM-8-288G B300-SXM, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- H100-SXM-8-80G H100-SXM, on-demand (×8 node): $0/vCPU-h, $0/GiB-h
- H100-SXM-4-80G H100-SXM, on-demand (×4 node): $0/vCPU-h, $0/GiB-h
- H100-SXM-2-80G H100-SXM, on-demand (×2 node): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, egress allowlist, static egress IP, private networking, SSO, GPU, Python, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Scaleway Instances (scaleway)
https://www.scaleway.com/en/pricing/virtual-instances/ · hyperscaler · isolation: vm
- shared-development: DEV1-L $0.048743352/h; DEV1-M $0.022979008799999995/h; DEV1-S $0.010212892799999998/h
- general-purpose: BASIC3-X12C-24G $0.30845758/h; BASIC3-X12C-48G $0.35317312/h; BASIC3-X16C-32G $0.4042000365999999/h
- dedicated-compute: COMPUTE3-X12C-24G $0.39948158/h; COMPUTE3-X16C-32G $0.5327179599999999/h; COMPUTE3-X24C-48G $0.79896316/h
- arm: BASIC2-A12C-24G $0.17954483999999998/h; BASIC2-A12C-48G $0.20560045999999998/h; BASIC2-A16C-32G $0.23518325999999998/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, static egress IP, private networking, SSO, GPU, arm64, Windows, live fork (no pause), audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## ScitiX Agent Sandbox (scitix-agent-sandbox)
https://scitix.github.io/Agent-Sandbox/ · agent-sandbox · isolation: container
- Self-hosted runtime + externally billed clusters
- Self-hosted container backend
- Self-hosted microVM backend
- Pre-warmed pool capacity
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, open source, GPU, arm64, Python, Node.js, Go, Rust, MCP server, inbound access rules, secret proxy, secret proxy for any API, extra volumes, shared volumes

## Scrapeless Agent Browser (scrapeless)
https://www.scrapeless.com/en/pricing · browser · isolation: not published
- Basic browser from-rate: Basic $0.09/h
- Growth browser from-rate: Growth $0.081/h
- Scale browser from-rate: Scale $0.076/h
- Business browser from-rate: Business $0.072/h
Features: persistent disk, idle auto-stop, browser, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, SOC 2, HIPAA, SSO, Python, Node.js, MCP server, hosted agent API (their own agent)

## Scrapfly Cloud Browser (scrapfly)
https://scrapfly.io/pricing · browser · isolation: not published
- Discovery overage: Discovery $0.018/h
- Pro overage: Pro $0.042/h
- Startup overage: Startup $0.024/h
- Enterprise overage: Enterprise $0.0144/h
Features: idle auto-stop, browser, desktop GUI, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, SOC 2, HIPAA, Python, Node.js, spend limits, MCP server, inbound access rules

## Scraping Bee (scrapingbee)
https://www.scrapingbee.com/pricing/ · browser · isolation: not published
- Browser API request (not time-priced): Browser API request (not time-priced) $null/h
Features: browser, anti-bot stealth, residential IPs, open internet, SOC 2, Python, Node.js, EU data residency, MCP server

## SeaCloudAI Sandbox (seacloudai)
https://sandbox-gateway.cloud.seaart.ai · agent-sandbox · isolation: container
- On-demand
Features: pause/resume, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, code interpreter, preinstalled agents, HTTPS preview URLs, egress allowlist, Python, Node.js, Go, wake on request, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Sealos DevBox (sealos-devbox)
https://sealos.io/pricing/ · dev-env · isolation: container
- Monthly resource package (flat, shared by all DevBoxes/apps in the plan)
- New-user price on first paid plan (Starter $7, Hobby $25)
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, root, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, self-hosting, Python, Node.js, Go, Rust, Java, spend limits, MCP server, automatic snapshots, snapshots on demand, agent harness API, inbound access rules

## Self-hosted bare-metal sandbox fleet (hardware floor) (self-host-baremetal)
https://www.hetzner.com/dedicated-rootserver/ax42/ · self-host · isolation: firecracker
- Hetzner AX42 / 50% occupancy: $0.017220898683078554/vCPU-h, $0.005535288862418106/GiB-h: 
- Hetzner AX42 / 70% occupancy: $0.012300641916484683/vCPU-h, $0.003953777758870076/GiB-h: 
- Hetzner AX42 / 90% occupancy: $0.00956716593504364/vCPU-h, $0.0030751604791211693/GiB-h: 
- Hetzner EX44 / 70% occupancy: $0.0055092316855271/vCPU-h, $0.0017708244703479963/GiB-h: 
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, self-hosting, BYOC, open source, arm64, memory fork, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Server4Agent (server4agent)
https://www.server4agent.com/pricing · dev-env · isolation: container
- Native product meter; not a general-purpose 4-vCPU/8-GiB VM: 
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, HTTPS preview URLs, open internet, BYOC, Python, Node.js, wake on request, webhooks, EU data residency, scoped API keys, spend limits, MCP server, agent harness API, hosted agent API (their own agent), inbound access rules

## SF Compute (GPU market + Autoresearch sandboxes) (sfcompute)
https://autoresearch.sfcompute.com/ · gpu-cloud · isolation: vm
- Autoresearch RL sandbox (billed per GiB of RAM, auto-park): sandbox-sm $0.09144/h; sandbox-md $0.73152/h; sandbox-lg $1.46304/h
- Autoresearch CPU node (cpu-2 / cpu-8 container, persistent disk): cpu-2 $0.18/h; cpu-8 $0.4824/h
- Autoresearch interactive H100 node, on demand (h100-1 = 1 GPU $3.96/h; h100-8 = 8 GPUs): $0/vCPU-h, $0/GiB-h
- Autoresearch batch jobs on H100 (queued, preemptible): $0/vCPU-h, $0/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, HTTPS preview URLs, egress allowlist, open internet, private networking, SOC 2, GPU, Python, Node.js, wake on request, live fork (no pause), memory fork, audit logs, EU data residency, scoped API keys, spend limits, MCP server, snapshots on demand, hosted agent API (their own agent), fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## SF Tensor (sf-tensor)
https://sf-tensor.com · gpu-cloud · isolation: gvisor
- Native product; see regime notes
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, BYOC, GPU, Python, gVisor or VM (no shared kernel), shared volumes

## Shadeform (shadeform)
https://www.shadeform.ai/ · gpu-cloud · isolation: dedicated-vm
- Native product; see regime notes
- H100 PCIe, on-demand marketplace (x1, from $2.50/h): $0/vCPU-h, $0/GiB-h
- L40S marketplace listing (1x): $0/vCPU-h, $0/GiB-h
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, SSH, public IPv4, open internet, BYOC, SOC 2, HIPAA, GPU, Python, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## Shardflux (shardflux)
https://shardflux.dev/#pricing · agent-sandbox · isolation: firecracker
- Workspace: $0.12/vCPU-h, $0.04/GiB-h
- File-first workspace (fresh VM per command): $0.12/vCPU-h, $0.04/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), root, browser, browser automation API, egress allowlist, open internet, Python, Node.js, wake on request, live fork (no pause), memory fork, audit logs, spend limits, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## shellbox (shellbox)
https://shellbox.dev/ · agent-sandbox · isolation: firecracker
- On-demand (Firecracker VM, default): x1 $0.02/h; x2 $0.04/h; x3 $0.06/h
- Cloud Hypervisor VM (--ch, nested virtualization): x1 $0.02/h; x2 $0.04/h; x3 $0.06/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, open internet, static egress IP, Python, wake on request, memory fork, scheduled runs (cron), EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables)

## Shuttle (shuttle)
https://www.shuttle.dev · paas · isolation: not published · discontinued

Features: ≥24 h sessions

## Sieve (sieve)
https://sievedata.com/ · gpu-cloud · isolation: not published · discontinued

Features: ≥24 h sessions

## Signadot (signadot)
https://www.signadot.com/ · dev-env · isolation: container
- Native product; see regime notes
Features: fork/clone, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, HTTPS preview URLs, custom domains, BYOC, SOC 2, SSO, arm64, MCP server, inbound access rules

## Skyhook (skyhook)
https://skyhook.io · paas · isolation: container
- Native service
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, HTTPS preview URLs, egress allowlist, BYOC, SOC 2, SSO, GPU, arm64, Python, Node.js, wake on request, live resize, EU data residency, inbound access rules, extra volumes

## Skyvern (skyvern)
https://www.skyvern.com/pricing · browser · isolation: container
- Managed browser service
Features: fork/clone, browser, desktop GUI, computer-use API, browser automation API, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, static egress IP, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, arm64, Python, Node.js, webhooks, audit logs, EU data residency, MCP server, snapshots on demand, hosted agent API (their own agent)

## smol machines (smol-machines)
https://smolmachines.com/pricing · agent-sandbox · isolation: vm
- Active usage plus running-instance fee: $0.05/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, open source, SSO, arm64, Python, Node.js, wake on request, live fork (no pause), memory fork, spend limits, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Smooth (smooth)
https://www.smooth.sh/pricing · browser · isolation: not published
- Managed browser service
Features: ≥24 h sessions, browser, computer-use API, browser automation API, anti-bot stealth, CAPTCHA solving, egress allowlist, open internet, BYOC, HIPAA, SSO, Python, Node.js, zero data retention, hosted agent API (their own agent)

## Solari (solari)
https://docs.getsolari.com/pricing · agent-sandbox · isolation: vm
- On-demand: $0.0245/vCPU-h, $0.0077/GiB-h
- Starter tier rate: $0.035/vCPU-h, $0.011/GiB-h
- Free tier rate: $0.0525/vCPU-h, $0.0165/GiB-h
- Enterprise published rate: $0.0175/vCPU-h, $0.0055/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, anti-bot stealth, CAPTCHA solving, HTTPS preview URLs, egress allowlist, HIPAA, Python, Node.js, wake on request, live fork (no pause), memory fork, spend limits, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Specific (specific)
https://specific.dev/pricing · paas · isolation: container
- Hosted services: Tiny $0.009/h; Small $0.025/h; Medium $0.12/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, HTTPS preview URLs, custom domains, self-hosting, BYOC, SSO, Python, Node.js, Go, Rust, Java, live resize, scheduled runs (cron), EU data residency, scoped API keys, inbound access rules, extra volumes

## Sprites (Fly.io) (sprites)
https://fly.io/pricing · agent-sandbox · isolation: firecracker
- Sprite, billed on actual CPU time + actual memory while awake: $0.03825/vCPU-h, $0.021875/GiB-h
Features: snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, SOC 2, HIPAA, SSO, Python, Node.js, Go, Rust, Java, credential injection, HTTP method/path egress rules, wake on request, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## StackBlitz WebContainers (webcontainers)
https://stackblitz.com/pricing · dev-env · isolation: v8-isolate
- Browser-local runtime licence: 
Features: snapshots, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, browser, HTTPS preview URLs, self-hosting, BYOC, SSO, Node.js, EU data residency, snapshots on demand, gVisor or VM (no shared kernel)

## STACKIT Compute Engine (stackit)
https://pim.api.stackit.cloud/v1/skus · hyperscaler · isolation: vm
- Single AZ: t2i.1 (shared vCPU) $0.01608/h; t3i.1 (dedicated vCPU) $0.028061/h; c2i.1 (shared vCPU) $0.024627/h
- Metro HA: t2i.1 (shared vCPU) $0.020018/h; t3i.1 (dedicated vCPU) $0.056121/h; c2i.1 (shared vCPU) $0.049255/h
- GPU Server (eu01): n2.14d.g1 (1x L40S 48GB) $1.983075/h; n2.28d.g2 (2x L40S) $3.96615/h; n2.56d.g4 (4x L40S) $7.932299/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, egress allowlist, static egress IP, private networking, SOC 2, SSO, GPU, arm64, Windows, live fork (no pause), EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## StarSling (starsling)
https://starsling.dev/pricing · dev-env · isolation: firecracker
- CPU CI runners: 2 $0.24/h; 4 $0.48/h; 8 $0.96/h
- RTX PRO 6000 GPU runner (private beta): RTX PRO 6000 $3.5532/h
- RTX 5090 GPU runner (private beta): RTX 5090 $1.8132/h
- RTX 4090 GPU runner (private beta): RTX 4090 $1.5132/h
Features: idle auto-stop, full VM (own kernel), Docker inside, SSO, GPU, Python, Node.js, gVisor or VM (no shared kernel), firewall inside (nftables)

## StateSet Sandbox (stateset)
https://sandbox.stateset.app/ · agent-sandbox · isolation: gvisor
- On-demand: $0.128/vCPU-h, $0.0106/GiB-h
- Team overage: $0.1/vCPU-h, $0.008/GiB-h
Features: snapshots, fork/clone, persistent disk, custom image (Docker or snapshot), start from your own snapshot, preinstalled agents, HTTPS preview URLs, egress allowlist, private networking, self-hosting, SOC 2, SSO, Python, Node.js, credential injection, webhooks, audit logs, spend limits, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, secret proxy

## Steel.dev (steel)
https://docs.steel.dev/overview/pricinglimits · browser · isolation: container
- Browser session, Launch rate ($0.10/browser-hour): browser-session $0.1/h
- Browser session, Scale rate ($0.08/browser-hour): browser-session $0.08/h
- Enterprise (custom rates, reserved browser pools, stealth browser): browser-session $null/h
Features: persistent disk, idle auto-stop, ≥24 h sessions, full VM (own kernel), browser, computer-use API, browser automation API, desktop control API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, open internet, static egress IP, self-hosting, open source, HIPAA, SSO, Python, Node.js, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel)

## Strong Compute (strong-compute)
https://strongcompute.com · finops (GPU/LLM spend-management SaaS; legacy managed GPU training platform 'ISC') · isolation: container
- Native product; see regime notes
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, open internet, SOC 2, HIPAA, SSO, GPU, Python, EU data residency, spend limits, automatic snapshots, snapshots on demand, extra volumes, shared volumes

## Superserve (superserve)
https://superserve.ai/pricing · agent-sandbox · isolation: firecracker
- Sandbox, pay-as-you-go (allocated vCPU + RAM per second while active): $0.0504/vCPU-h, $0.0162/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, code interpreter, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, SOC 2, HIPAA, Python, Node.js, credential injection, wake on request, memory fork, audit logs, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API

## Surfsky (surfsky)
https://surfsky.io/pricing · browser · isolation: container
- PAYG browser: PAYG $0.35/h
- Starter browser: Starter $0.3/h
- Base browser: Base $0.25/h
- Advanced browser: Advanced $0.2/h
Features: snapshots, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, browser, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, egress allowlist, open internet, Python, Node.js, EU data residency

## Tabstack (tabstack)
https://tabstack.ai/ · browser · isolation: not published
- Native product meter; not a general-purpose 4-vCPU/8-GiB VM: 
Features: ≥24 h sessions, browser, Python, Node.js, spend limits, hosted agent API (their own agent)

## tama (tama)
https://tama.computer/ · agent-sandbox · isolation: vm
- CPU on-demand: $0.01166/vCPU-h, $0.005523/GiB-h
- RTX 4090 live bundle: RTX 4090 $0.54/h
- A 6000 live bundle: A 6000 $0.675/h
- RTX 5090 live bundle: RTX 5090 $0.8775/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, desktop GUI, preinstalled agents, SSH, HTTPS preview URLs, GPU, Python, Node.js, live fork (no pause), automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Tart + Orchard (Cirrus Labs) (tart-orchard)
https://tart.run/licensing/ · macos · isolation: apple-vm
- Self-hosted licence on your own Macs (compute not included): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, root, desktop GUI, SSH, egress allowlist, self-hosting, BYOC, open source, arm64, macOS, live fork (no pause), automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Teclada (teclada)
https://www.teclada.com/ · dev-env · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, SSH, self-hosting, BYOC, SSO, audit logs, inbound access rules

## Tencent Cloud Agent Runtime — Agent Sandbox (tencent-agent-runtime)
https://cloud.tencent.com/document/product/1814/133249 · agent-sandbox · isolation: vm
- Agent Sandbox allocated runtime: $0.043454327326/vCPU-h, $0.013411829422/GiB-h
- Elastic Deployment beta
Features: snapshots, memory snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, HTTPS preview URLs, egress allowlist, private networking, SOC 2, HIPAA, SSO, Windows, Python, Node.js, HTTP method/path egress rules, wake on request, EU data residency, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Tencent Cloud CubeSandbox (tencent-cubesandbox)
https://github.com/TencentCloud/CubeSandbox · agent-sandbox · isolation: vm
- On-demand
- Self-hosted CubeSandbox (Apache-2.0)
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, browser, desktop GUI, preinstalled agents, HTTPS preview URLs, egress allowlist, open internet, static egress IP, self-hosting, open source, arm64, Python, Node.js, credential injection, HTTP method/path egress rules, wake on request, live fork (no pause), memory fork, EU data residency, automatic snapshots, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Tencent Cloud SCF (tencent-scf)
https://cloud.tencent.com/document/product/583/17299 · hyperscaler · isolation: container
- Event function
- Web function
- Provisioned concurrency (idle): $0/vCPU-h, $0.030492/GiB-h
Features: volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, EU data residency, inbound access rules, extra volumes, shared volumes

## Tencent Cloud Studio (tencent-cloud-studio)
https://cloud.tencent.cn/document/product/1039/131894 · dev-env · isolation: container
- IDE pay-as-you-go: 1C2G $0.014902032691/h; 2C4G $0.037255081727/h; 4C8G $0.074510163453/h
Features: fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), Docker inside, preinstalled agents, HTTPS preview URLs, GPU, Python, Node.js, wake on request, live resize, inbound access rules

## Tenki Sandbox (tenki)
https://tenki.cloud/pricing · agent-sandbox · isolation: vm
- Sandbox session (allocated vCPU + RAM per second while running): $0.0504/vCPU-h, $0.0162/GiB-h
- Tenki Runners, Linux x64 GitHub Actions (per core-minute): tenki-standard-small-2c-4g $0.24/h; tenki-standard-medium-4c-8g $0.48/h; tenki-standard-large-8c-16g $0.96/h
- Tenki Runners, macOS (Apple M4 Pro) GitHub Actions (per core-minute): tenki-macos-26-mini $2.4/h; tenki-macos-26-small $4.8/h; tenki-macos-26-medium $7.2/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, static egress IP, SOC 2, arm64, macOS, Python, Node.js, Go, memory fork, spend limits, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## TensorDock (tensordock)
https://www.tensordock.com/ · gpu-cloud · isolation: vm
- H100 SXM5 on-demand (marketplace floor): $0.003/vCPU-h, $0.002/GiB-h
- A100 SXM4, VRAM unverified, on-demand
- RTX 4090 on-demand (marketplace floor): $0.003/vCPU-h, $0.002/GiB-h
- H100, spot
Features: pause/resume, persistent disk, ≥24 h sessions, full VM (own kernel), Docker inside, root, desktop GUI, SSH, public IPv4, raw TCP inbound, open internet, static egress IP, GPU, Windows, Python, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Tensorlake Sandboxes (tensorlake)
https://www.tensorlake.ai/pricing · agent-sandbox · isolation: firecracker
- Pro rate set (active CPU, allocated RAM/disk): $0.042/vCPU-h, $0.009/GiB-h
- BYOC / Self-hosted Compute (your AWS bill, reference rate, + 30% fee): $0.094375/vCPU-h, $0/GiB-h
- Usage Credits prepaid rate set: $0.07/vCPU-h, $0.015/GiB-h
- Free unmetered sandbox: Free $0/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, systemd, browser, desktop GUI, computer-use API, browser automation API, desktop control API, browser + desktop control, SSH, HTTPS preview URLs, egress allowlist, open internet, private networking, BYOC, SOC 2, HIPAA, SSO, GPU, Python, Node.js, wake on request, live fork (no pause), memory fork, webhooks, EU data residency, zero data retention, spend limits, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## TensorPool (tensorpool)
https://tensorpool.dev · gpu-cloud · isolation: not published · discontinued

Features: volumes, ≥24 h sessions, GPU, firewall inside (nftables), extra volumes, shared volumes

## Texel.ai (texel-ai)
https://texel.ai · gpu-cloud · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, GPU, webhooks

## Thunder Compute (thunder-compute)
https://www.thundercompute.com/pricing · gpu-cloud · isolation: container
- GPU instances (on-demand, per GPU incl. base vCPU/RAM): $0/vCPU-h, $0/GiB-h
- GPU sandbox (Firecracker, per second, TTL): $0.051/vCPU-h, $0.016/GiB-h
- Reserved clusters (8-512+ GPUs, fixed term)
Features: snapshots, fork/clone, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, Docker inside, SSH, public IPv4, HTTPS preview URLs, egress allowlist, open internet, self-hosting, BYOC, SOC 2, GPU, Python, live resize, live fork (no pause), MCP server, snapshots on demand, inbound access rules, firewall inside (nftables)

## Tilde.run (discontinued) (tilde-run)
https://lakefs.io/blog/we-recently-shut-down-tilde-run/ · agent-sandbox · isolation: container · discontinued

Features: snapshots, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, root, egress allowlist, open internet, HTTP method/path egress rules, audit logs, extra volumes, shared volumes

## Tilion (tilion)
https://tilion.com/pricing · browser · isolation: not published
- Managed browser service
Features: idle auto-stop, browser, desktop GUI, computer-use API, browser automation API, anti-bot stealth, residential IPs, open internet, self-hosting, open source, MCP server, hosted agent API (their own agent), inbound access rules

## Tinfoil (tinfoil)
https://tinfoil.sh · paas · isolation: vm
- Native product; see regime notes: $0.06/vCPU-h, $0.01/GiB-h
- Confidential GPU enclave (monthly contract): $0.06/vCPU-h, $0.01/GiB-h
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, open source, SOC 2, HIPAA, SSO, GPU, Python, Node.js, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## TinyFish (tinyfish)
https://www.tinyfish.ai/pricing · browser · isolation: not published
- Native product meter; not a general-purpose 4-vCPU/8-GiB VM: 
- Monitor (Page/Topic change tracking): 
Features: idle auto-stop, ≥24 h sessions, browser, browser automation API, anti-bot stealth, residential IPs, open internet, BYOC, SSO, Python, Node.js, webhooks, scheduled runs (cron), audit logs, MCP server, hosted agent API (their own agent)

## Together AI (together-gpu)
https://www.together.ai/pricing · gpu-cloud · isolation: container
- H100 GPU cluster, on-demand (8-GPU nodes, per GPU-hour): $0/vCPU-h, $0/GiB-h
- H200 HGX/SXM, on-demand (×8 node)
- B200 HGX/SXM, on-demand (×8 node)
- B300 HGX/SXM, on-demand (×8 node)
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, Docker inside, nested virtualization, SSH, public IPv4, HTTPS preview URLs, static egress IP, SOC 2, HIPAA, SSO, GPU, Python, Node.js, EU data residency, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Together Code Sandbox (together-code-sandbox)
https://www.together.ai/pricing · agent-sandbox · isolation: firecracker
- Code Sandbox VMs on a Together custom plan (sales only; self-serve = codesandbox-sdk card): $0.0446/vCPU-h, $0.0149/GiB-h
- Together Code Interpreter ($0.03 per 60-minute session): session-60min $0.03/h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, code interpreter, HTTPS preview URLs, custom domains, open internet, SOC 2, SSO, Python, Node.js, Go, Rust, wake on request, live resize, live fork (no pause), memory fork, EU data residency, spend limits, MCP server, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Trainy (trainy)
https://trainy.ai/ · paas · isolation: container
- Native product; see regime notes
Features: pause/resume, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, SSH, public IPv4, HTTPS preview URLs, self-hosting, BYOC, open source, SSO, GPU, Python, credential injection, wake on request, audit logs, EU data residency, MCP server, inbound access rules

## Trigger.dev (trigger-dev)
https://trigger.dev/pricing · paas · isolation: container
- Managed tasks: Micro $0.060840000000000005/h; Small 1x $0.12168000000000001/h; Small 2x $0.243/h
Features: snapshots, memory snapshots, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, browser, open internet, static egress IP, private networking, self-hosting, open source, SOC 2, HIPAA, SSO, Python, Node.js, wake on request, webhooks, scheduled runs (cron), EU data residency, scoped API keys, spend limits, MCP server, automatic snapshots, hosted agent API (their own agent)

## Ubicloud (ubicloud)
https://www.ubicloud.com/docs/about/pricing · hyperscaler · isolation: vm
- Dedicated VMs Germany: standard-2 $0.0624/h; standard-4 $0.1248/h; standard-8 $0.2496/h
- Shared burstable VMs: burstable-1 $0.0156/h; burstable-2 $0.0312/h
- Standard x64 CI: 2 $0.075/h; 4 $0.15/h; 8 $0.3/h
- Premium x64 CI: 2 $0.12/h; 4 $0.24/h; 8 $0.48/h
Features: snapshots, fork/clone, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, SSH, public IPv4, open internet, static egress IP, private networking, self-hosting, BYOC, open source, SOC 2, SSO, GPU, arm64, audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## UCloud Agent Sandbox (ucloud)
https://astraflow.ucloud.cn/docs/agent-sandbox · agent-sandbox · isolation: firecracker
- On-demand (UCloud China account, CNY list price): $0.016094/vCPU-h, $0.008047/GiB-h
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, desktop GUI, computer-use API, code interpreter, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, Python, Node.js, credential injection, wake on request, live fork (no pause), memory fork, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Unikraft Cloud (unikraft-cloud)
https://unikraft.com/pricing · agent-sandbox · isolation: firecracker
- Self-serve plan pool (Hobby / Team / Pro concurrent running capacity)
- Enterprise: dedicated hosts / on-prem / BYOC (custom price)
Features: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), nested virtualization, root, HTTPS preview URLs, custom domains, egress allowlist, private networking, self-hosting, BYOC, SOC 2, HIPAA, GPU, arm64, Python, Node.js, credential injection, wake on request, live resize, live fork (no pause), memory fork, scheduled runs (cron), audit logs, EU data residency, automatic snapshots, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## UpCloud (upcloud)
https://upcloud.com/pricing/ · hyperscaler · isolation: vm
- Starter: starter-1c-1g $0.005208333333333333/h; starter-1c-2g $0.010416666666666666/h; starter-2c-2g $0.01488095238095238/h
- Premium: premium-1c-1g $0.008928571428571428/h; premium-1c-2g $0.017857142857142856/h; premium-2c-2g $0.026785714285714284/h
- Cloud-Native: cloud-native-1c-4g $0.020833333333333332/h; cloud-native-2c-4g $0.026785714285714284/h; cloud-native-1c-8g $0.03422619047619048/h
- GPU Server (on-demand, Helsinki): $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SSO, GPU, Windows, live resize, audit logs, EU data residency, spend limits, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Upstash Box (upstash-box)
https://upstash.com/pricing/box · agent-sandbox · isolation: container
- Pay as You Go, Small box (2 vCPU / 4 GB), active CPU core-hours: $0.1/vCPU-h, $0/GiB-h
- Pay as You Go, Medium box (4 vCPU / 8 GB), active CPU core-hours: $0.2/vCPU-h, $0/GiB-h
- Pay as You Go, Large box (8 vCPU / 16 GB), active CPU core-hours: $0.4/vCPU-h, $0/GiB-h
- Fixed / Keep-Alive Small box, always on ($8/box/month): small $0.010959/h
Features: snapshots, fork/clone, pause/resume, persistent disk, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, browser, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, SOC 2, arm64, Python, Node.js, Go, Rust, credential injection, wake on request, live fork (no pause), webhooks, scheduled runs (cron), MCP server, snapshots on demand, agent harness API, hosted agent API (their own agent), inbound access rules, secret proxy, secret proxy for any API

## use.computer (use-computer)
https://use.computer/ · macos · isolation: apple-vm
- Reserved M4 Mac mini, priced per macOS VM slot (2 VMs per Mac): split: macOS VM slot = half of a reserved M4 Mac mini (5 CPU / 8 GiB) $0.45/h; whole: one VM per reserved Mac (10 CPU / 16 GiB) $0.9/h
Features: snapshots, fork/clone, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), browser, desktop GUI, computer-use API, desktop control API, browser + desktop control, SSH, static egress IP, arm64, macOS, Python, Node.js, live fork (no pause), snapshots on demand, gVisor or VM (no shared kernel), inbound access rules

## Vast.ai (vast-ai)
https://github.com/vast-ai/docs/blob/main/guides/pricing.mdx · gpu-cloud · isolation: container
- RTX4090, on-demand
- RTX5090, on-demand
- RTX-PRO-6000, on-demand
- A100-80GB, on-demand
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, browser, desktop GUI, residential IPs, SSH, HTTPS preview URLs, open internet, static egress IP, SOC 2, HIPAA, GPU, arm64, Python, EU data residency, snapshots on demand, inbound access rules, firewall inside (nftables), extra volumes

## Veertu Anka Build Cloud (anka-build-cloud)
https://docs.veertu.com/anka/licensing/ · macos · isolation: apple-vm
- Anka Build licence on your own Macs (contact sales): Anka Build host licence (price on request) $null/h
- Anka Build AMI on AWS EC2 Mac (Marketplace hourly software fee): Basic mac2.metal (M1) $0.28/h; Basic mac2-m2.metal $0.28/h; Basic mac-m4.metal $0.35/h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, desktop GUI, preinstalled agents, SSH, egress allowlist, open internet, static egress IP, self-hosting, BYOC, SSO, arm64, macOS, MCP server, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Vercel Sandbox (vercel-sandbox)
https://vercel.com/docs/sandbox/pricing · agent-sandbox · isolation: firecracker
- On-demand, iad1 / cle1 / pdx1 (base rate): $0.128/vCPU-h, $0.0212/GiB-h
- Region pinned: bom1 (Mumbai): $0.14/vCPU-h, $0.0232/GiB-h
- Region pinned: yul1 (Montreal): $0.147/vCPU-h, $0.0243/GiB-h
- Region pinned: sin1 (Singapore) / arn1 (Stockholm): $0.16/vCPU-h, $0.0266/GiB-h
Features: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, preinstalled agents, SSH, HTTPS preview URLs, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, Python, Node.js, credential injection, HTTP method/path egress rules, live resize, live fork (no pause), EU data residency, spend limits, MCP server, automatic snapshots, snapshots on demand, fast boot (benchmarked), gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

## Verda (formerly DataCrunch) (verda)
https://verda.com/pricing · gpu-cloud · isolation: vm
- 1x H100 SXM5 80GB on-demand: $0/vCPU-h, $0/GiB-h
- 1x H100 SXM5 80GB spot: $0/vCPU-h, $0/GiB-h
- 1x H200 SXM5 141GB on-demand: $0/vCPU-h, $0/GiB-h
- 1x H200 SXM5 141GB spot: $0/vCPU-h, $0/GiB-h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, static egress IP, SOC 2, GPU, Python, wake on request, audit logs, EU data residency, spend limits, MCP server, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables), extra volumes, shared volumes

## Vibrant Labs (vibrant-labs)
https://vibrantlabs.com/ · agent-sandbox · isolation: not published
- Native service
Features: ≥24 h sessions

## Volcano Engine AgentKit / veFaaS sandbox (volcengine-agentkit)
https://docs.volcengine.com/docs/agentkit/Billing_items?lang=zh · agent-sandbox · isolation: vm
- AgentKit Sandbox tool (mainland China, cn-beijing): $0.021727163663/vCPU-h, $0.005472026404/GiB-h
- Agent application Runtime (hosts the agent app; mainland China, cn-beijing): $0.052239075597/vCPU-h, $0.008291729422/GiB-h
- Direct veFaaS sandbox tariff not verified (mainland China): $0.021727163663/vCPU-h, $0.005472026404/GiB-h
- AgentKit Sandbox on BytePlus (Asia Pacific, Johor ap-southeast-1), USD list: $0.0351/vCPU-h, $0.0081/GiB-h
Features: snapshots, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, HTTPS preview URLs, egress allowlist, Python, Node.js, wake on request, MCP server, automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), inbound access rules, secret proxy, extra volumes, shared volumes

## Voltage Park (voltage-park)
https://www.voltagepark.com/pricing · gpu-cloud · isolation: vm
- HGX H100 on-demand (Ethernet), from $1.99/GPU-h: $0/vCPU-h, $0/GiB-h
- HGX H100 on-demand InfiniBand (8+ GPUs): $0/vCPU-h, $0/GiB-h
Features: pause/resume, persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, root, SSH, public IPv4, open internet, SOC 2, HIPAA, GPU, Python, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Vultr Cloud Compute (vultr)
https://www.vultr.com/pricing/ · hyperscaler · isolation: vm
- Cloud Compute Regular (vc2, shared vCPU), hourly with 672 h monthly cap: vc2-1c-1gb (Cloud Compute Regular, shared) $0.0074/h; vc2-1c-2gb (Cloud Compute Regular, shared) $0.0149/h; vc2-2c-2gb (Cloud Compute Regular, shared) $0.0223/h
- Cloud Compute High Performance (vhp, AMD EPYC or Intel Xeon, NVMe, shared vCPU), hourly with 672 h cap: vhp-1c-1gb-amd (High Performance AMD, shared) $0.0089/h; vhp-1c-2gb-amd (High Performance AMD, shared) $0.0179/h; vhp-2c-2gb-amd (High Performance AMD, shared) $0.0268/h
- Cloud Compute High Frequency (vhf, 3 GHz+ Intel, NVMe, shared vCPU), hourly with 672 h cap: vhf-1c-1gb (High Frequency Intel 3GHz+, shared) $0.0089/h; vhf-1c-2gb (High Frequency Intel 3GHz+, shared) $0.0179/h; vhf-2c-2gb (High Frequency Intel 3GHz+, shared) $0.0268/h
- Optimized Cloud Compute (voc, dedicated AMD EPYC vCPU: General / CPU / Memory / Storage Optimized), hourly with 672 h cap: voc-c-1c-2gb-25s-amd (Optimized Cloud Compute CPU Optimized, dedicated) $0.0417/h; voc-g-1c-4gb-30s-amd (Optimized Cloud Compute General Purpose, dedicated) $0.0446/h; voc-m-1c-8gb-50s-amd (Optimized Cloud Compute Memory Optimized, dedicated) $0.0595/h
Features: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, open internet, static egress IP, private networking, SSO, GPU, Windows, audit logs, EU data residency, snapshots on demand, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Vultr Cloud Compute (Windows Server) (vultr-windows)
https://docs.vultr.com/support/platform/billing/is-a-windows-license-included-in-the-monthly-price · windows · isolation: vm
- Cloud Compute (Regular Performance) VM with Windows Server, licence included: vc2-1c-2gb + Windows licence: 1 vCPU, 2 GB, 55 GB SSD ($10 base + $16 licence = $26/mo) $0.03869/h; vc2-2c-2gb + Windows licence: 2 vCPU, 2 GB, 65 GB SSD ($15 base + $32 licence = $47/mo) $0.06994/h; vc2-2c-4gb + Windows licence: 2 vCPU, 4 GB, 80 GB SSD ($20 base + $32 licence = $52/mo) $0.077381/h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, root, SSH, public IPv4, raw TCP inbound, open internet, static egress IP, private networking, SSO, Windows, audit logs, EU data residency, snapshots on demand, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## Vultr Cloud GPU (vultr-gpu)
https://api.vultr.com/v2/plans?type=vcg · gpu-cloud · isolation: vm
- vcg-a16-2c-8g-2vram, on-demand (×1): $0/vCPU-h, $0/GiB-h
- vcg-a16-2c-16g-4vram, on-demand (×1): $0/vCPU-h, $0/GiB-h
- vcg-a40-1c-5g-2vram, on-demand (×1): $0/vCPU-h, $0/GiB-h
- vcg-a16-3c-32g-8vram, on-demand (×1): $0/vCPU-h, $0/GiB-h
Features: snapshots, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, open internet, static egress IP, SOC 2, HIPAA, SSO, GPU, Python, Node.js, live resize, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

## WarpBuild (GitHub Actions runners) (warpbuild)
https://www.warpbuild.com/pricing · macos · isolation: vm
- macOS runner, Apple M4 Pro: 6 vCPU, 14 GB ($0.08/min) $4.8/h; 12 vCPU, 28 GB ($0.16/min) $9.6/h
- Windows runner, x86-64: 4 vCPU, 16 GB ($0.016/min) $0.96/h; 8 vCPU, 32 GB ($0.032/min) $1.92/h; 16 vCPU, 64 GB ($0.064/min) $3.84/h
- Linux runner, x86-64: 2 vCPU, 8 GB ($0.004/min) $0.24/h; 4 vCPU, 16 GB ($0.008/min) $0.48/h; 8 vCPU, 32 GB ($0.016/min) $0.96/h
- Linux runner, arm64: 2 vCPU, 8 GB ($0.003/min) $0.18/h; 4 vCPU, 16 GB ($0.006/min) $0.36/h; 8 vCPU, 32 GB ($0.012/min) $0.72/h
Features: snapshots, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, preinstalled agents, SSH, egress allowlist, open internet, static egress IP, private networking, BYOC, SOC 2, SSO, arm64, Windows, macOS, Python, Node.js, wake on request, EU data residency, scoped API keys, MCP server, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Wasmer (wasmer)
https://wasmer.io/pricing · paas · isolation: isolate
- WebAssembly Edge: $0.1/vCPU-h, $0.01/GiB-h
Features: pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, SSH, public IPv4, HTTPS preview URLs, custom domains, raw TCP inbound, open internet, self-hosting, open source, Python, Node.js, wake on request, EU data residency, MCP server, gVisor or VM (no shared kernel), extra volumes, shared volumes

## webapp.io (webapp-io)
https://webapp.io · dev-env · isolation: not published · discontinued

Features: ≥24 h sessions, automatic snapshots, snapshots on demand, inbound access rules, firewall inside (nftables)

## Windmill (windmill)
https://www.windmill.dev/pricing · paas · isolation: container
- Cloud Team execution credits
- Dedicated Cloud Enterprise
- Self-hosted software
Features: persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, browser, preinstalled agents, HTTPS preview URLs, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, arm64, Python, Node.js, audit logs, EU data residency, MCP server, agent harness API, inbound access rules, extra volumes

## Windows 365 Cloud PC (Business / Enterprise) (windows-365)
https://www.microsoft.com/en-us/windows-365/business/compare-plans-pricing · windows · isolation: vm
- Windows 365 Business (per user per month): Business 2 vCPU, 8 GB, 128 GB $0.049315/h; Business 4 vCPU, 16 GB, 128 GB $0.076712/h; Business 8 vCPU, 32 GB, 256 GB $0.149041/h
- Windows 365 Enterprise (per user per month): Enterprise 2 vCPU, 8 GB, 128 GB $0.056164/h; Enterprise 4 vCPU, 16 GB, 128 GB $0.090411/h; Enterprise 8 vCPU, 32 GB, 256 GB $0.180822/h
- Windows 365 Enterprise GPU Standard: GPU Standard $0.735616/h
- Windows 365 Flex: 
Features: snapshots, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, nested virtualization, root, browser, desktop GUI, egress allowlist, open internet, private networking, SOC 2, HIPAA, SSO, GPU, Windows, Python, Node.js, audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables)

## Windows 365 for Agents (windows-365-agents)
https://learn.microsoft.com/en-us/windows-365/agents/pricing-paygo-always-available · windows · isolation: vm
- Pay-as-you-go Cloud PC for Agents (US geography): Cloud PC for Agents (hardware spec not published) $0.4/h
Features: idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), full VM (own kernel), browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, egress allowlist, private networking, SOC 2, HIPAA, SSO, Windows, HTTP method/path egress rules, audit logs, EU data residency, MCP server, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables)

## YepCode Run (yepcode)
https://yepcode.io/pricing/ · paas · isolation: firecracker
- Process execution on Starter (per-second Yeps): Starter process sandbox (300 MB RAM, vCPU not published) $0.040961/h
- Process execution on Growth (per-second Yeps): Growth process sandbox (600 MB RAM, vCPU not published) $0.02048/h
Features: full VM (own kernel), code interpreter, HTTPS preview URLs, static egress IP, self-hosting, BYOC, SOC 2, SSO, Python, Node.js, scheduled runs (cron), audit logs, EU data residency, MCP server, gVisor or VM (no shared kernel), inbound access rules

## Zeabur (zeabur)
https://zeabur.com/pricing · paas · isolation: container
- Bring or buy server; compute separately billed: Tencent Lighthouse 2C2G (Silicon Valley) $0.00411/h; Tencent Lighthouse 2C4G (Silicon Valley) $0.00548/h; Tencent Lighthouse 2C8G (Silicon Valley) $0.00822/h
Features: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, SSH, public IPv4, HTTPS preview URLs, custom domains, static egress IP, BYOC, GPU, Python, Node.js, EU data residency, spend limits, MCP server, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes

## Zenrows Browser Sessions (zenrows)
https://www.zenrows.com/pricing · browser · isolation: container
- Build B1 top-up equivalent: Build B1 $0.17727272727272728/h
- Build B2 top-up equivalent: Build B2 $0.15/h
- Build B3 top-up equivalent: Build B3 $0.13604651162790699/h
- Launch L1 top-up equivalent: Launch L1 $0.117/h
Features: idle auto-stop, browser, computer-use API, browser automation API, browser + desktop control, anti-bot stealth, residential IPs, open internet, SOC 2, SSO, Python, Node.js, MCP server

## Zhipu Z Managed Agents (zhipu)
https://docs.bigmodel.cn/cn/managed-agents/overview.md · agent-sandbox · isolation: not published
- Managed Agent sandbox, temporary free runtime fee
Features: snapshots, pause/resume, persistent disk, volumes, idle auto-stop, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, root, browser, egress allowlist, open internet, Python, Node.js, Go, Rust, Java, credential injection, wake on request, scheduled runs (cron), MCP server, automatic snapshots, agent harness API, hosted agent API (their own agent), secret proxy, secret proxy for any API, shared volumes

## Zibra Labs (zibra-labs)
https://zibralabs.ai/ · gpu-cloud · isolation: not published
- Native product; see regime notes
Features: ≥24 h sessions, GPU

## zipbox (zipbox)
https://zipbox.ai/pricing · agent-sandbox · isolation: firecracker
- Per-second fixed VM sizes: Tiny $0.0068/h; Small $0.0137/h; Medium $0.0274/h
Features: snapshots, pause/resume, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, full VM (own kernel), Docker inside, root, browser, desktop GUI, computer-use API, browser + desktop control, preinstalled agents, SSH, HTTPS preview URLs, Python, Node.js, credential injection, live resize, snapshots on demand, gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy, secret proxy for any API

## Zo Computer (zo-computer)
https://www.zo.computer/pricing · dev-env · isolation: container
- Zo Basic subscription, one always-on computer (4 cores / 32 GB) - $18/mo: basic $0.024658/h
- Zo Pro subscription, one always-on computer (16 cores / 128 GB) - $64/mo: pro $0.087671/h
- Zo Ultra subscription, one always-on computer (64 cores / 512 GB) - $200/mo: ultra $0.273973/h
- N Basic workspaces sized to peak concurrency (part-time use)
Features: snapshots, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, root, browser, browser automation API, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, Python, Node.js, scheduled runs (cron), MCP server, hosted agent API (their own agent), inbound access rules