boat vs zipbox
Both give a coding agent its own Linux virtual machine with root, a browser and keys kept outside the box. zipbox is a personal cloud computer you run from a dashboard, while boat is built to be driven by code: an API for sandboxes and agents, forks, snapshots and bigger machines.
Where boat stands out
Run sandboxes and agents from code
boat has an API to create, start, stop, fork and delete sandboxes, and to prompt, steer and follow the agent inside each one. You can drive agents inside from outside, from your backend, a queue or a CI job. zipbox is used through its dashboard, browser terminal and SSH, and documents no public API.
Fork, snapshots and a disk that keeps up
boat snapshots the disk every minute and on stop, and can fork a running sandbox so many agents start from the same prepared state. zipbox saves the disk when you pause a box, and has no clone or fork.
A bigger, fuller machine
boat machines go up to 16 vCPU / 32 GB and expose /dev/kvm, so Android emulators and VMs run inside. You also get raw TCP, a dedicated IP and HTTPS on up to 50 ports, token-protected by default.
The companies and products
boat
boat.dev runs sandboxes for AI agents: each one is a full Linux VM with its own kernel.
You start a machine with one command or one API call, then SSH in, run Docker, open a desktop, or let a built-in agent (Claude Code, Codex, OpenCode and others) work in it. The disk stays between sessions, and a stopped sandbox costs nothing.
- Isolation
- Full VM per sandbox, own kernel
- Regions
- EU: Germany, Finland, France
zipbox
zipbox, made by Tribes, Inc., calls itself a cloud computer for your AI agent: a Linux machine where Claude Code, Codex or another agent runs on its own.
Each zipbox box is a Firecracker microVM with its own kernel, root access, an X11 desktop with Chrome that you can watch and take over, an email address, a public domain and two crypto wallets. You create and use boxes from a web dashboard, a browser terminal or SSH, pick one of 13 agents at boot, and pay from prepaid credit by the second. It is made for one person who wants an agent working on its own machine, not for running fleets from code.
Which use cases each one fits
For each use case, Battleships prices every sandbox provider on the same workload and sorts them from the cheapest monthly bill up. “3rd cheapest of 367” means 2 of the other 366 providers run that workload, with every feature it needs, for less. “With limits” means the provider runs it but misses something, so it is listed after every full match. Open a use case for its whole ranking and what each provider misses.
| Use case and its workload | boat: price rank | zipbox: price rank |
|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A software factory starts, forks and drives many agents from code, which needs an API that boat has and zipbox does not document. | 3rd cheapest of 367 $634 a month | 21st of 367, with limits $482 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB Both fit an assistant that lives on its own machine all month; zipbox adds an email inbox and wallets, boat adds snapshots every minute and an API. | 5th cheapest of 366 $657 a month | 3rd cheapest of 366 $500 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month | 186th of 345, with limits $2,000 a month | 52nd of 345, with limits $23 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Giving a sandbox to each of your users means creating them from your backend, and zipbox has no public API for that or a published limit on how many can run. | Cheapest of 361 $209 a month | 33rd of 361, with limits $159 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 34th of 360, with limits $2,000 a month | 8th of 360, with limits $685 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month Evals fork one prepared machine into many copies, which boat does from a running sandbox and zipbox does not offer. | 10th cheapest of 364 $200 a month | 94th of 364, with limits $137 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI runners are started by a pipeline, not by hand, and builds of Android apps or VM images need nested virtualization, which boat has and zipbox does not. | 4th cheapest of 377 $86 a month | 225th of 377, with limits $88 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB Both fit a personal dev box; zipbox is cheaper at the smallest sizes, boat gives a dedicated IP, raw TCP and machines up to 16 vCPU. | 18th cheapest of 368 $26 a month | 11th cheapest of 368 $20 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Both put a real desktop and browser on the machine that you can watch and take over; boat streams it at 60fps and has a computer-use API you call from outside. | 6th cheapest of 369 $200 a month | 4th cheapest of 369 $137 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month boat drives Chromium from an API at scale, while zipbox's browser is used by the agent on that one box. | 7th cheapest of 367 $200 a month | 51st of 367, with limits $69 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month | 6th cheapest of 346 $90 a month | 2nd cheapest of 346 $34 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month | 71st of 338, with limits $1,800 a month | 38th of 338, with limits $680 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 2nd cheapest of 360 $200 a month | 13th of 360, with limits $34 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month | Can't run it | Can't run it |
| macOS agents 5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month | Can't run it | Can't run it |
Key features
| Feature | boat | zipbox |
|---|---|---|
| Isolation | Full VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts | Firecracker microVM with its own kernel under KVM |
| Control from code | REST API to create, start, stop, fork and delete sandboxes | Web dashboard, browser terminal and SSH; no public API documented |
| Coding agents | Seven agents preinstalled, with a prompt, steer and events API | 13 agents to pick from at boot, used through the browser terminal |
| Biggest machine | Up to 16 vCPU / 32 GB on the $100 plan and above | Up to 6 vCPU / 16 GB, bigger sizes through support |
| Saving state | Disk snapshots every minute and on stop; a stopped sandbox keeps its disk for free | Pause archives the disk for free; restore boots it on fresh hardware |
| Fork | Fork a running sandbox with its disk | Not offered; the machine menu has no clone or fork |
| Nested virtualization | Yes, /dev/kvm for Android emulators and VMs | Not offered; Firecracker does not expose /dev/kvm to guests |
| Root and Docker | Root, Docker and systemd | Root inside the box; Docker not documented but runs in a microVM with its own kernel |
| Inbound access | HTTPS URLs for up to 50 ports, token-protected by default; raw TCP; dedicated IPv4 or IPv6 | HTTPS subdomains through Caddy, 30 records per box; SSH over IPv6 only |
| Keeping API keys out of the sandbox | Credentials proxy for model providers and any HTTPS API; switch a credential to another account when credits run out | Placeholder keys swapped at the network edge for included and user-added hosts, up to 8 hosts per key |
| Browser and desktop | 60fps Linux desktop with Chromium and a computer-use API | X11 desktop with Chrome that you watch and take over in the dashboard; the agent drives it through a skill inside the box |
| Email and wallets | No built-in mailbox or wallet | Each box has an email address and custodial EVM and Solana wallets |
| Smallest machine | From 2 vCPU / 4 GB | Tiny 1 vCPU / 2 GB for light agents |
| Regions | EU only (Germany, Finland, France) | Not published |
| Compliance and transparency | SOC 2 is in progress. We're open about where we stand: live status and data retention policy. | No SOC 2 report published |
Every feature Battleships tracks (43)
| Feature | boat | zipbox |
|---|---|---|
| Full VM (own kernel) | Yes | Yes |
| Docker inside | Yes | Yes |
| Nested virtualization | Yes | No |
| Root | Yes | Yes |
| Systemd | Yes | Not published |
| Desktop GUI | Yes | Yes |
| Browser | Yes | Yes |
| Computer-use API | Yes | Yes |
| SSH | Yes | Yes |
| Persistent disk | Yes | Yes |
| Pause/resume | Yes | Yes |
| Automatic snapshots | Yes | No |
| Fork/clone | Yes | No |
| Live fork (no pause) | Yes | No |
| Memory fork | No | No |
| Custom image (Docker or snapshot) | Yes | Yes |
| ≥24 h sessions | Yes | Yes |
| Idle auto-stop | No | No |
| Wake on request | No | No |
| HTTPS preview URLs | Yes | Yes |
| Raw TCP inbound | Yes | Not published |
| Public IPv4 | No | No |
| Egress allowlist | No | No |
| Secret proxy | Yes | Yes |
| Secret proxy for any API | Yes | Yes |
| Preinstalled agents | Yes | Yes |
| Agent harness API | Yes | No |
| MCP server | No | No |
| GPU | No | No |
| Windows | No | No |
| Arm64 | No | No |
| Self-hosting | No | No |
| BYOC | No | No |
| Open source | No | No |
| Scoped API keys | Yes | Not published |
| Spend limits | Yes | Not published |
| Usage per sandbox (API) | Yes | No |
| Audit logs | No | Not published |
| SSO | No | No |
| SOC 2 | No | No |
| HIPAA | No | No |
| EU data residency | Yes | No |
| Zero data retention | Yes | Not published |
Performance
There is no comparable measurement for zipbox. boat did 26.69 runs per second on the StarSling hpc-sandbox-benchmarks Node.js build-and-test loop at 4 vCPU / 8 GB, on dedicated AMD Ryzen 9 9950X hosts, but zipbox has not been run on the same test.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. zipbox has no published run in this suite yet, so there is no like-for-like number to compare.
Pricing in five scenarios
Each bill is priced from the providers' public price lists by the open Battleships engine, with the cheapest plan that runs the scenario.
| Scenario | boat, a month | zipbox, a month |
|---|---|---|
| One coding agent, workdays 4 vCPU / 8 GB, 50 GB disk, 8 hours a day, 22 days a month, files kept between days | $20 · $20/mo | $5 · Prepaid pay as you go |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $26 · $20/mo | $20 · Prepaid pay as you go |
| 50 agents in parallel 50 machines of 4 vCPU / 8 GB at once, 8-hour sessions, 1,100 sessions a month, Docker inside | $317 · $20/mo × 3 seats | $241 · Prepaid pay as you go |
| Sandboxes for your users 2 vCPU / 4 GB, 30,000 sessions of 10 minutes a month, up to 300 at once | $200 · $100/mo × 2 seats | $69 · Prepaid pay as you go · with limits |
| 1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month | $2,000 · $2000/mo · with limits | $137 · Prepaid pay as you go · with limits |
boat: Plans start at $20 a month, and the whole fee is usage credit. You pay per second while a sandbox runs, and a stopped sandbox costs nothing and keeps its disk.
zipbox: zipbox bills per second from prepaid credit, with no subscription, at a fixed hourly rate per size. Paused boxes cost nothing, and model calls through zipbox are billed at provider cost plus 3%.
Limits in those scenarios
zipbox does not publish how many boxes can run at once, so the scenarios with 300 or 1,000 sandboxes are not confirmed to fit, and it publishes no egress price. Machines stop at 6 vCPU / 16 GB without asking support, and when credit hits zero, running boxes pause after one hour. boat's limits are published per plan; at 1,000 short jobs at once, even the $2,000 plan queues starts.
| Scenario | boat | zipbox |
|---|---|---|
| One coding agent, workdays | None that bite |
|
| One agent that never stops | None that bite |
|
| 50 agents in parallel | None that bite |
|
| Sandboxes for your users | None that bite |
|
| 1,000 short jobs at once |
|
|
Which one should you use?
Use zipbox when
you are one person who wants a cheap cloud computer for an agent, run from a dashboard, with its own email inbox and crypto wallets.
- An email address and wallets on every box. Each zipbox box comes with its own inbox and crypto wallets, useful for agents that sign up for services or move funds.
- Small machines and prepaid credit. A Tiny 1 vCPU box and prepaid credit with no subscription suit one person running a light agent now and then.
- More agents to pick from. zipbox offers 13 agents at boot, including Grok, Hermes, OpenClaw, Cline, Cursor CLI and Amp.
Use boat when
your code starts and steers the agents: fleets for your users, evals or CI, forks of a prepared machine, emulators, and machines up to 16 vCPU.
- Run sandboxes and agents from code. Without an API, every new box and every agent task on zipbox starts with a person clicking in the dashboard, so a fleet for your users, evals or CI cannot be started by your code.
- Fork, snapshots and a disk that keeps up. Without fork, each zipbox box has to install the same repo and tools from scratch, and work done since the last pause is not saved anywhere else.
- A bigger, fuller machine. zipbox tops out at 6 vCPU / 16 GB without asking support, Firecracker gives no nested virtualization for emulators, and SSH reaches boxes only over IPv6.
Questions people ask
Is boat cheaper than zipbox?
No. On list price zipbox costs less in every scenario on this page, for example $241 against $317 a month for the agent fleet. boat's fee is usage credit and buys published concurrency limits, an API and forks, which the larger scenarios rely on.
Does zipbox have an API?
zipbox documents a web dashboard, a browser terminal and SSH, plus command line tools inside the box for email, wallets and domains. It documents no public API to create boxes or drive agents. boat has one for both.
Can zipbox fork or snapshot a box?
zipbox can pause a box, which archives its disk for free, and restore it on fresh hardware without its running processes. It has no clone or fork. boat saves the disk every minute and can fork a running sandbox.
Can I run Docker and Android emulators in each?
Both give you root in a VM with its own kernel, so Docker works in both, though zipbox does not document it. Android emulators need /dev/kvm, which boat exposes and Firecracker microVMs like zipbox's do not.
Where does zipbox run?
zipbox does not publish its regions. boat runs in the EU only, in Germany, Finland and France.
Is boat SOC 2 compliant?
SOC 2 is in progress. We're open about where we stand: live status at status.ascii.dev and our data retention policy at docs.boat.dev/data-retention.
Start a sandbox
Install the CLI, then run boat new.
7-day free trial