# HUD

Official pricing: https://www.hud.ai/  
Category: agent-sandbox · Isolation: vm

## Pricing regimes (raw)

- **Native service** (sizes): Micro (t3.small) 2 vCPU/2 GiB $null/h; Small (t3.medium) 2 vCPU/4 GiB $null/h; Default (t3.large) 2 vCPU/8 GiB $null/h; Large (t3.xlarge) 4 vCPU/16 GiB $null/h; Extra Large (t3.2xlarge) 8 vCPU/32 GiB $null/h [alt]
- **HUD-hosted on Modal (CPU/GPU profiles)** (sizes): cpu=1;memory_mb=8192 1 vCPU/8 GiB $null/h; cpu=2;memory_mb=8192 2 vCPU/8 GiB $null/h; cpu=4;memory_mb=8192 4 vCPU/8 GiB $null/h; cpu=16;memory_mb=16384 16 vCPU/16 GiB $null/h; cpu=8;memory_mb=32768;gpu=L4:1 8 vCPU/32 GiB $null/h; cpu=8;memory_mb=32768;gpu=H100!:1 8 vCPU/32 GiB $null/h [alt]

## Features

Yes: ≥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

No: snapshots, memory snapshots, fork/clone, pause/resume, persistent disk, volumes, idle auto-stop, start from your own snapshot, full VM (own kernel), nested virtualization, anti-bot stealth, CAPTCHA solving, residential IPs, public IPv4, HTTPS preview URLs, static egress IP, arm64, Windows, macOS, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, automatic snapshots, snapshots on demand, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy for any API, extra volumes, shared volumes

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

## Caveats

- Official website pricing: SDK/platform free; Cloud $0.10/environment-hour, 100+ parallel instances, $10 free credits. Enterprise custom adds 24h runtime and training; academic grant $100 conditional. CPU/RAM allocation unspecified, so environment-hour not a 4/8 VM guarantee.
- Null values mean unknown, not free/unlimited. No model-token cost, tax or unverified ancillary charge included.
- Native browser/device/task/credit prices with unknown CPU/RAM are deliberately non-priceable. Check regime before interpreting any partial resource estimate.
- Docs name Free / Pro / Scale / Enterprise subscription tiers (team storage caps 10 GiB / 50 GiB / 200 GiB / unlimited); Pro and Scale fees are not public (hud.ai/pricing redirects to sign-in).

## How this provider charges


YC: Winter 2025; directory status **Active**. Research scope: public-product-researched. Source: https://www.ycombinator.com/companies/hud

HUD RL environments

## Regime table

| Regime | When it applies | How billed | Numbers | Source |
|---|---|---|---|---|
| Native service | Hardware and complete resource tariff unverified. | resource; CPU alloc, memory alloc | Unknown or native-unit billing; see facts below | https://www.hud.ai/ |
| Custom / unverified | Plan limits apply | Monthly fee; credits only as stated | fee=None; included usage=$None; concurrent=None; max session h=None | https://www.hud.ai/ |
| Retained disk / GiB-month | Lifecycle/scope must be checked | Native add-on | unverified, not $0 | https://www.hud.ai/ |
| Snapshots / GiB-month | Lifecycle/scope must be checked | Native add-on | unverified, not $0 | https://www.hud.ai/ |
| Egress / GiB | Lifecycle/scope must be checked | Native add-on | unverified, not $0 | https://www.hud.ai/ |
| IPv4 / month | Lifecycle/scope must be checked | Native add-on | unverified, not $0 | https://www.hud.ai/ |

## Verified billing facts and gotchas

1. Official website pricing: SDK/platform free; Cloud $0.10/environment-hour, 100+ parallel instances, $10 free credits. Enterprise custom adds 24h runtime and training; academic grant $100 conditional. CPU/RAM allocation unspecified, so environment-hour not a 4/8 VM guarantee.

Native pricing (not coerced into incompatible engine units):
```json
{
  "environment_hour_usd": 0.1
}
```

## Worked example

Requested: 4 vCPU/8 GiB, 50 simultaneous × 8 h/day × 22 days = **8,800 instance-hours**, 30% CPU; 50 GiB snapshots and 100 GiB egress. A month is 730 hours only when explicitly used in rate conversion.

8,800 environment-hours ×$0.10=$880 before enterprise runtime limits or other charges. CPU/RAM shape is unspecified; 8h sessions may need Enterprise because the website reserves extended 24h runtime for it. Not a 4/8 price quote.

## Sources and coverage

- https://www.ycombinator.com/companies/hud
- https://www.hud.ai/

Page captures and documentation indexes/bundles are in `raw/`; provider JSON lists the evidence paths. HN query results were captured separately as `<id>--hn.txt`; they are discovery leads, not current tariff authority. Downloaded docs do not imply every feature was verified; unsupported assertions remain null.