# shellbox

Official pricing: https://shellbox.dev/  
Category: agent-sandbox · Isolation: firecracker

## Pricing regimes (raw)

- **On-demand (Firecracker VM, default)** (sizes): x1 2 vCPU/4 GiB $0.02/h; x2 4 vCPU/8 GiB $0.04/h; x3 6 vCPU/12 GiB $0.06/h; x4 8 vCPU/16 GiB $0.08/h; x5 10 vCPU/20 GiB $0.1/h; x6 12 vCPU/24 GiB $0.12/h
- **Cloud Hypervisor VM (--ch, nested virtualization)** (sizes): x1 2 vCPU/4 GiB $0.02/h; x2 4 vCPU/8 GiB $0.04/h; x3 6 vCPU/12 GiB $0.06/h; x4 8 vCPU/16 GiB $0.08/h; x5 10 vCPU/20 GiB $0.1/h; x6 12 vCPU/24 GiB $0.12/h

## Features

Yes: 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)

No: volumes, browser, desktop GUI, computer-use API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, public IPv4, egress allowlist, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, arm64, Windows, macOS, Node.js, HTTP method/path egress rules, live resize, live fork (no pause), MCP server, agent harness API, hosted agent API (their own agent), inbound access rules, secret proxy, secret proxy for any API, extra volumes, shared volumes

Unknown: everything else. Evidence (source + quote) per feature: https://battleships.dev/data/providers/shellbox.json → feature_evidence

## Caveats

- Public list price per x1 slot. Running and parked are mutually exclusive time states. SSH key identifies account; prepaid $10 minimum top-up. Sizes scale linearly x1 to x8.
- 16 running slots and 64 total slots per account; a box counts as its size in slots, so a 4 vCPU/8 GB (x2) box allows 8 concurrent. Encoded as plan concurrency 16 (x1 bound) + max_total_vcpu 32; the engine does not yet enforce slot-weighted quota (see research/verify/engine-requests.md).
- Features set to null were not verified. No sandbox was purchased or tested; feature/performance statements are documentation claims. Null maximums are unknown unless a cited note explicitly states unlimited.
- A complete monthly bill requires known snapshot/egress/plan limits; do not silently treat missing ancillary charges as free.
- Parked fee $0.50/x1 slot-month is per slot, not per GiB; snapshot_gib_month 0.01 = $0.50 / 50 GB assumes the retained state equals the box disk size. Parked storage is otherwise free ("memory snapshots ... on NVMe, rootfs images ... on HDDs"). No desktop/GUI/browser image is documented (desktop/browser/computer_use=false). (shellbox.dev, 2026-09-28)

## How this provider charges


As of 2026-09-28. Prices are USD unless explicitly labelled otherwise. Source type: official published list/promo or unknown, never a fabricated quote.

Public list price per x1 slot. Running and parked are mutually exclusive time states. SSH key identifies account; prepaid $10 minimum top-up. Sizes scale linearly x1 to x8.

| Regime | When it applies | How billed / numbers | Source |
|---|---|---|---|
| Running | x1 slot = 2 vCPU / 4 GB / 50 GB | $0.02/slot-hour, by the minute; x1–x8. | https://shellbox.dev/ |
| Parked | Stopped box retains disk | $0.50/slot-month. | https://shellbox.dev/ |
| Always-on | keepalive enabled | Running metering continues; no published monthly compute cap. | https://shellbox.dev/ |

## Gotchas

- Public list price per x1 slot. Running and parked are mutually exclusive time states. SSH key identifies account; prepaid $10 minimum top-up. Sizes scale linearly x1 to x8.
- 16 running slots and64 total slots per account. 4/8 box=x2, so at most8 concurrently under default quota. Engine does not express slot-weighted quota.
- Features set to null were not verified. No sandbox was purchased or tested; feature/performance statements are documentation claims. Null maximums are unknown unless a cited note explicitly states unlimited.
- A complete monthly bill requires known snapshot/egress/plan limits; do not silently treat missing ancillary charges as free.

## Worked workload

4 vCPU / 8 GiB; 50 simultaneous instances × 8 h/day × 22 days = **8,800 instance-hours**, **35,200 vCPU-hours**, **70,400 GiB-hours**. At 30% observed CPU:10,560 active-vCPU-hours only where the meter actually uses utilization. Assume one start/instance/day:1,100starts. Retain50GiB snapshots for a month;100GiB outbound. Active disk capacity was not specified.

Choose x2 (4 vCPU/8 GB/100 GB): 8,800×$0.04 = $352 running compute. Each of 50 instances runs 176h. If $1/x2-month parked price prorates across remaining 554/730 month, parked cost is $37.95; proration must be verified, so do not treat this as a firm total. Snapshot/extra egress charges and 50-way account capacity are unverified. Official full FAQ resolves capacity: C=50 x2 boxes require100running slots, exceeding16running and64total slot caps. Workload infeasible on default account.

## Sources

- https://shellbox.dev/