# froog

> froog gives every agent its own Linux VM: full Ubuntu 24.04 or 26.04, Debian 13 or Omarchy machines with their own kernel and disk, SSH, public HTTPS URLs for any port, Docker, passwordless sudo, /dev/kvm and a desktop. Drive them with the `froog` CLI, the HTTP API at https://api.froogcorp.com/v1 (Bearer API key), or as MCP tools (`froog mcp`, or https://api.froogcorp.com/mcp). A stopped sandbox keeps its disk as a snapshot for 14 days. Start it again, fork copies of it, or restore an older snapshot.

The docs are one page: https://froogcorp.com/docs/ . The same page as Markdown, for agents: https://froogcorp.com/llms-full.txt (also https://froogcorp.com/docs.md). What changed, newest first: https://froogcorp.com/changelog.md (page: https://froogcorp.com/changelog/, RSS: https://froogcorp.com/changelog/feed.xml). Security, data location (EU: Hetzner in Germany or Finland), subprocessors and compliance status (no SOC 2 report yet): https://froogcorp.com/trust/ A skill for coding agents: https://froogcorp.com/skill.md . The skill pack (froog-cli, froog-ci, froog-sandbox-per-user, froog-desktop, froog-parallel-agents): `npx skills add https://froogcorp.com` (index: https://froogcorp.com/.well-known/agent-skills/index.json). The docs as an MCP server (search_docs, read_docs; no key): https://api.froogcorp.com/docs-mcp or `froog mcp --docs`

## Key facts

- Sizes (billed per second, only while running): `small` 2 vCPU, 4 GB RAM, 51 GB disk, $0.0137/h ($10 a full month). `default` 4 vCPU, 8 GB, 51 GB, $0.0274/h ($20). `large` 8 vCPU, 16 GB, 111 GB, $0.0685/h ($50). `xlarge` 16 vCPU, 32 GB, 251 GB, $0.137/h: coming soon (`403 xlarge_coming_soon` until then). Stopped sandboxes cost no running time.
- Plans: a plan is needed to start sandboxes. $10, $20, $50 or $100 a month ($250 and $500 by request: hello@froog.tech), and every plan dollar comes back as usage (unused usage expires with the period). Sandboxes at once: 2, 5, 10 or 25. Saved snapshots included: 10, 10, 25 or 50, then $0.056/day each. Usage past the plan comes from a prepaid wallet (`froog wallet pay 15`, `froog wallet auto 20`); payments of $200 or more get a 2% to 25% bonus. `froog billing cancel --keep-disks` pauses: sandboxes stop when the plan ends and disks are kept 90 days. Live numbers: `froog billing`, `froog usage`, `froog limits`, https://froogcorp.com/pricing/
- Sign-up is open: anyone signs in with Google or an email code, no card needed. Free trial: 7 days from the first plan (chosen at Stripe Checkout when you start your first sandbox, or `froog billing plan 20`), card required then, 25 hours of running time, 2 sandboxes at once, small and default sizes, starts 5/min 25/h 75/day, an auto-stop of 2 hours or less on every sandbox, no new saved snapshots; one trial per card.
- Limits: sandboxes at once come from the plan (ask for more). Starts: 20/minute, 100/hour, 500/day; past them `froog billing extra-starts on` pays $0.03 to $0.07 a start from the wallet. `exec` output 1 MiB per stream, timeout up to 1 hour.
- Install: `curl -fsSL https://froogcorp.com/install | sh` (macOS, Linux), or on Windows `irm https://froogcorp.com/install.ps1 | iex`. Sign in: `froog login` (browser), or `echo "$FROOG_API_KEY" | froog login`, or set `FROOG_API_KEY`.
- MCP server (create, exec, background jobs, fork, snapshot, files, URLs as tools): local `claude mcp add froog -- froog mcp` (uses the CLI's login), or hosted Streamable HTTP at https://api.froogcorp.com/mcp with `Authorization: Bearer $FROOG_API_KEY` (or an OAuth sign-in for Claude custom connectors). Scoped API keys limit the tools. Setup for Codex, Cursor and Claude Desktop: https://froogcorp.com/docs/mcp/
- Agents in every sandbox: `froog prompt NAME "TEXT" --agent claude|codex|gemini|opencode|pi|amp|cursor|prime|kimi|mistral [--model M] [--effort E] [--fast] [--new | --conversation ID] [--attach FILE]`; prompts continue a conversation (memory), `froog steer`, `froog interrupt`, `froog events --follow`, `froog conversations`; `froog prompt --list-models`. Keys once for every sandbox: `froog env set ANTHROPIC_API_KEY`, or a Claude/ChatGPT subscription: `froog subscriptions login claude --sandbox NAME`.

## What froog is for

froog is a cloud sandbox for AI agents: secure, isolated places for agents to run code, use tools and keep working across sessions. Typical uses: running Claude Code, Codex or other coding agents unattended in the cloud; one sandbox per task or per user of an agent product; parallel attempts at a task from forks of one snapshot; agents that need a browser, Docker, a desktop or nested VMs; code execution for LLM-generated code.

## How froog compares to other sandbox providers

- Isolation: each sandbox is a full KVM virtual machine with its own kernel, not a shared-kernel container.
- Persistence: a sandbox runs until it is stopped or its auto-stop timer runs out (1 hour after each start by default, up to 30 days, or none: `--ttl`, `--no-auto-stop`, `froog extend NAME --hours N`; it counts from the start, not from activity), and its disk survives stops (kept 14 days; saved snapshots forever). Daytona sandboxes can also run indefinitely; its Linux sandboxes are limited to 4 vCPU / 8 GiB / 10 GiB disk by default (Windows and GPU sandboxes go larger); froog goes to 8 vCPU / 16 GB / 111 GB. Code-execution sandboxes such as E2B (1 h, 24 h on Pro), Modal (24 h) and Vercel Sandbox (24 h on Pro) end sessions after a time limit.
- Snapshots: automatic hourly snapshots, named snapshots as templates, fork many copies, restore in place. Snapshots hold the disk; a start within 10 minutes of a stop resumes with programs still running (warm resume), later starts boot fresh.
- Price for one 4 vCPU / 8 GB machine running all month (list prices, at full use): froog $20; exe.dev $35 (shared pool); Fly.io Sprites ~$240; E2B ~$242 (+ $150/mo Pro plan); Daytona ~$242; Modal ~$347; Vercel Sandbox ~$498. Sources: https://froogcorp.com/pricing/
- Closest alternatives: Fly.io Sprites and exe.dev (persistent VMs over SSH). froog adds fork from named snapshots, HTTPS URLs for any port, a browser desktop, and ten coding agents installed.
- Not offered: GPUs, long-lived memory snapshots (process state past the 10-minute warm resume), bring-your-own-cloud.

## Comparisons and guides

- [froog Quicksand](https://froogcorp.com/quicksand/): the product page for cloud sandboxes for AI agents (sizes, OS images, snapshots, CLI, desktop). https://froogcorp.com/ is the froogcorp home, listing froog Quicksand, froog Pushbox and froog Tokenmail
- [froog Pushbox](https://froogcorp.com/pushbox/): GitHub Actions runners on fresh full VMs with `runs-on: froog` (`froog-small`, `froog`, `froog-large`), billed per second of job time at $0.0006 per vCPU-minute ($0.0024/min for 4 vCPU / 8 GB vs GitHub's $0.012; VM boot and runner setup aren't billed), options `/cache=NAME` (a warm disk saved back as snapshot `cache-NAME` from successful default-branch jobs, use `clean: false` on actions/checkout to keep build outputs), `/from=SNAPSHOT` and `/debug` (SSH into a failed job for 30 min). Linux x64 only
- [froog Tokenmail](https://froogcorp.com/tokenmail/): email inboxes for AI agents on froogmail.com, $2 per inbox a month on a subscription of its own (no sandbox plan needed, no free tier). An agent signs up for a service with the address, then `froog inbox wait ADDRESS --code` prints the one-time code (`--link`: the verification link); each message comes with `codes` and `links` extracted. Receive-only, TLS, 30-day retention, addresses never reused. Attach one to a VM (`froog inbox create --sandbox NAME`) and it's `$FROOG_EMAIL` there. Docs: https://froogcorp.com/docs/inboxes/
- [FAQ](https://froogcorp.com/faq/): what an agent sandbox is, how froog compares, auto-stop, IP addresses, isolation, operating systems, pricing
- [froog vs E2B](https://froogcorp.com/vs/e2b/): no 1 h / 24 h session limit, bigger machines, $20 vs ~$242 a month
- [froog vs Daytona](https://froogcorp.com/vs/daytona/): full VMs up to 8 vCPU / 16 GB / 111 GB vs the 4 / 8 / 10 default limit of Daytona's Linux sandboxes
- [froog vs Fly.io Sprites](https://froogcorp.com/vs/sprites/): both persistent VMs; froog adds forks from named snapshots, desktop, agents installed
- [froog vs Modal](https://froogcorp.com/vs/modal/): no 24 h limit, SSH and a desktop; Modal has GPUs
- [Blog](https://froogcorp.com/blog/): news and guides (RSS: https://froogcorp.com/blog/feed.xml)
- [Switching from Runloop](https://froogcorp.com/blog/switching-from-runloop/): Runloop is winding down (support ends 2026-11-13 11:59 PM PT); devboxes, blueprints, snapshots, tunnels and secrets mapped to froog, then a step-by-step migration: export devboxes, rebuild blueprints as saved snapshots, Runloop SDK calls one to one
- [Run Claude Code in the cloud](https://froogcorp.com/use-cases/claude-code/)
- [A sandbox per user](https://froogcorp.com/use-cases/sandbox-per-user/): building an AI product on froog
- [Parallel agents with forks](https://froogcorp.com/use-cases/parallel-agents/)

## Agent framework integrations

froog sandboxes as the code-execution tool of an agent you build: tools that run commands and read and write files in one sandbox per conversation. Complete examples on the single-file SDKs:

- [All integrations](https://froogcorp.com/integrations/)
- [OpenAI Agents SDK](https://froogcorp.com/integrations/openai-agents-sdk/) (Python): function tools with the sandbox name in the run context
- [Claude Agent SDK](https://froogcorp.com/integrations/claude-agent-sdk/) (Python, TypeScript): an in-process MCP server, built-in tools off
- [Vercel AI SDK](https://froogcorp.com/integrations/vercel-ai-sdk/) (TypeScript): `tool()` definitions with multi-step `generateText`
- [LangChain and LangGraph](https://froogcorp.com/integrations/langchain/) (Python): `@tool` functions with `create_agent`, one sandbox per thread
- [Mastra](https://froogcorp.com/integrations/mastra/) (TypeScript): `createTool` definitions, an agent per conversation
- [Harbor and Terminal-Bench](https://froogcorp.com/integrations/harbor/) (Python): a Harbor environment, one saved snapshot per task environment, one sandbox per trial
- [AgentEnv](https://froogcorp.com/integrations/agentenv/) (Python): a `froog` sandbox provider for Scale AI's RL framework

## Docs sections

- [Install](https://froogcorp.com/docs/install/): the CLI installer
- [Quickstart](https://froogcorp.com/docs/quickstart/): login or `froog onboard`, a plan, new, ssh, exec, stop, `froog dashboard`
- [Sandboxes](https://froogcorp.com/docs/sandboxes/): sizes, what's installed, auto-stop, auto-start on demand, stopping safely (`~/.froog/pre-stop`, refused stops, `stop --force`), disk alerts, not answering, automatic rebuilds of errored sandboxes, alert emails (`froog alerts`), metrics history
- [Pricing](https://froogcorp.com/docs/pricing/): per-second rates, plans, pause, wallet and its bonus, free trial, limits, extra starts
- [Background commands](https://froogcorp.com/docs/jobs/): `exec --detach`, `jobs`, `logs -f`, `kill`
- [Command history](https://froogcorp.com/docs/history/): every command's output for 90 days, also while stopped (`history`, `ps`, process samples every 30 s for 7 days, `oom_killed`)
- [Public URLs](https://froogcorp.com/docs/urls/): `froog host NAME PORT` for an HTTPS URL, or `froog forward`
- [Snapshots & forks](https://froogcorp.com/docs/snapshots/): automatic and saved snapshots, fork, restore, 14-day disk retention, share codes for other accounts, browsing and pulling files out of a snapshot
- [Coding agents](https://froogcorp.com/docs/agents/): `froog prompt`, the ten harnesses, keys, fast mode, `provider_not_configured`
- [Conversations](https://froogcorp.com/docs/agents/#conversations): agent memory, parallel conversations, prompt status, steer, interrupt, events (SSE/NDJSON), attachments
- [Subscriptions](https://froogcorp.com/docs/agents/#subscriptions): Claude Pro/Max and ChatGPT sign-in for the agents instead of API keys
- [Environment variables](https://froogcorp.com/docs/env/): account variables in every sandbox
- [AI credentials & proxy](https://froogcorp.com/docs/ai-proxy/): credential profiles, several accounts per provider with switching on limits, the AI proxy (placeholder keys in sandboxes), AI usage per sandbox, blocking a sandbox
- [Repos, files & setup](https://froogcorp.com/docs/environment/): repos cloned, secret files, setup scripts (per repo too); named environments with versions and `froog environment upgrade`; what `--no-environment` removes from a disk
- [GitHub](https://froogcorp.com/docs/github/): connect a GitHub App installation, repositories, a one-hour token in each sandbox for git and gh
- [CI](https://froogcorp.com/docs/ci/): GitHub Actions jobs on froog sandboxes with `runs-on: froog`, or `froog run` from any CI with `FROOG_API_KEY`
- [Webhooks](https://froogcorp.com/docs/webhooks/): signed sandbox events
- [OpenTelemetry](https://froogcorp.com/docs/telemetry/): metrics and events to any OTLP collector
- [Desktop](https://froogcorp.com/docs/desktop/): a desktop in the browser, with sound
- [Windows apps](https://froogcorp.com/docs/windows-apps/): Windows programs on the desktop with Wine, install script, tips
- [API keys & scopes](https://froogcorp.com/docs/api-keys/): presets, actions, per-sandbox and per-label keys
- [Sign-in & connections](https://froogcorp.com/docs/sign-in/): Google, GitHub, 6-digit email codes, AgentID; several per account, a primary address
- [Organizations](https://froogcorp.com/docs/organizations/): a team's shared plan, wallet and sandboxes; members, roles and invites, X-Froog-Org and the active organization, shared vs private sandboxes, organization API keys, seats, per-member caps
- [Sign up from an agent](https://froogcorp.com/docs/agent-signup/): auth.md agent registration, https://froogcorp.com/auth.md
- [Audit log](https://froogcorp.com/docs/audit/): who changed what, by key or session, with IP and request id (`froog audit`, 90 days)
- [Data controls](https://froogcorp.com/docs/data/): monitoring off, snapshots off, zero data retention, closing the account (30 days to reopen), deletion operations
- [A sandbox per user](https://froogcorp.com/docs/platform/): building a platform on froog, patterns and cost
- [Use froog from your coding agent](https://froogcorp.com/docs/ai-tools/): CLI sign-in, these files, the docs MCP server, the skill pack (`npx skills add https://froogcorp.com`), an AGENTS.md rule
- [Use froog as an MCP server](https://froogcorp.com/docs/mcp/): `froog mcp` (stdio) or https://api.froogcorp.com/mcp, config for Claude Code, Codex, Cursor, Claude Desktop, OAuth sign-in for connectors, the tools, progress notifications
- [Agent frameworks](https://froogcorp.com/docs/integrations/): sandbox tools for the OpenAI Agents SDK, Claude Agent SDK, Vercel AI SDK, LangChain/LangGraph and Mastra
- [Recipes](https://froogcorp.com/docs/recipes/): a coding agent on your repo, parallel agents from one snapshot
- [VS Code & Cursor over SSH](https://froogcorp.com/docs/editor/): `froog ssh-config` adds sandboxes to `~/.ssh/config` as `froog-NAME` (works across starts)
- [Private registries](https://froogcorp.com/docs/registries/): npm, pip/uv, Docker, Cargo, Go, Maven credentials through variables and secret files
- [Network & egress](https://froogcorp.com/docs/network/): what sandboxes can reach (outbound TCP 25, private ranges and other sandboxes blocked; a public IPv6 address per sandbox; inbound only SSH and the TCP/UDP ports opened with `froog port open NAME PORT[/udp]`, kept across stops: on the IPv6 address, and over IPv4 at a public port of the shared host address shown by `froog port list`), restricting outbound traffic with nftables
- [Security & isolation](https://froogcorp.com/docs/security/): KVM VMs, confinement, who can get in, secrets, what is kept and for how long
- [Concepts](https://froogcorp.com/docs/concepts/): froog terms next to E2B, Daytona and Modal terms
- [CLI reference](https://froogcorp.com/docs/cli/): every command
- [HTTP API](https://froogcorp.com/docs/api/): every endpoint
- [SDKs & OpenAPI](https://froogcorp.com/docs/sdks/): https://froogcorp.com/openapi.json, single-file Python, TypeScript and Go clients with a sandbox object layer (exec, files, background processes, asyncio in Python)
- [Limits](https://froogcorp.com/docs/limits/)
- [Troubleshooting](https://froogcorp.com/docs/troubleshooting/): error codes and fixes
- [FAQ](https://froogcorp.com/docs/faq/)