boat vs Opensteer

Opensteer sells a team of ready-made browser agents that run on its cloud runtime or on your own machine. boat gives your agent a whole Linux VM with a browser and desktop inside, where it can also run code, Docker and servers.

Where boat stands out

A whole machine around the browser

A boat sandbox is a full Linux VM with root, SSH, Docker, systemd and nested virtualization, with Chromium and a 60fps desktop already inside. Your agent can browse a site, run a script on what it found, build an app and serve it on a public HTTPS URL, all on one machine.

State you can save, stop and fork

boat saves the whole disk every minute and when a sandbox stops, and a stopped sandbox costs nothing. You can fork a running sandbox to try several paths at once from the same state.

Any coding agent, driven from outside, with keys kept out

Claude Code, Codex and five other agents come preinstalled, and you prompt, steer and follow them from outside through one API. A credentials proxy keeps real API keys outside the sandbox for any HTTPS API, and a credential can switch to another account when one runs out of credits.

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.

Opensteer

Opensteer is a Y Combinator startup that makes specialized AI agents that browse the web, use your tools and run on schedules, in its cloud or on your machine.

Opensteer is a different kind of product from boat: an agent product with a browser runtime, not a general sandbox. Its open-source runtime gives coding agents like Codex and Claude Code control of a local Chrome or Edge, or of an Opensteer Cloud browser, over CDP. Opensteer Cloud adds hosted workspaces that run around the clock, with schedules, browser profiles, security controls and team access, billed per runtime-hour.

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 workloadboat: price rankOpensteer: price rank
Software factory
100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month
A software factory needs Docker, a shell and a place to build and serve apps, which Opensteer's hosted workspaces do not document.
3rd cheapest of 367
$634 a month
246th of 367, with limits
$3,515 a month
Always-on AI employees
50 machines 24/7 · 2 vCPU / 4 GB
Both can keep an agent running all month; boat also gives it root, Docker and hosting on the same machine.
5th cheapest of 366
$657 a month
79th of 366, with limits
$7,295 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
290th of 345, with limits
$662 a month
Sandboxes for your users
300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month
Giving each of your users a sandbox needs a published machine size and concurrency, which Opensteer does not list.
Cheapest of 361
$209 a month
293rd of 361, with limits
$2,320 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
285th of 360, with limits
$9,995 a month
Agent evals
250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month
Evals need many identical machines; boat can fork a running sandbox, and Opensteer documents no fork or snapshot.
10th cheapest of 364
$200 a month
281st of 364, with limits
$995 a month
CI runners
40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month
4th cheapest of 377
$86 a month
260th of 377, with limits
$235 a month
24/7 dev boxes
1 machine 24/7 · 4 vCPU / 8 GB
18th cheapest of 368
$26 a month
272nd of 368, with limits
$141 a month
Computer-use agents
50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month
Both have a computer-use runtime; boat streams a full Linux desktop at 60fps, while Opensteer does not publish what its computer runtime contains.
6th cheapest of 369
$200 a month
22nd cheapest of 369
$995 a month
Browser-use agents
50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month
Opensteer is built around browser agents and can also drive your own local Chrome; boat puts Chromium on a whole machine where the agent can also run code.
7th cheapest of 367
$200 a month
36th cheapest of 367
$995 a month
Scraping & bots (startup scale)
50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month
For startup-scale bots, boat gives each one a whole machine with a browser and a disk that is saved every minute.
6th cheapest of 346
$90 a month
73rd cheapest of 346
$995 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
286th of 338, with limits
$19,995 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
71st of 360, with limits
$495 a month
GPU eval jobs
10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month
Can't run itCan't run it
macOS agents
5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month
Can't run itCan't run it

Key features

FeatureboatOpensteer
What you getA full Linux VM with its own kernel and root, on dedicated AMD Ryzen 9 9950X hostsHosted agent workspaces with browser and computer runtime
Shell accessSSH, plus a command and file APIA remote toolbelt that runs commands and reads, writes and patches files; no SSH
Docker and nested virtualizationDocker, systemd and /dev/kvm for Android emulators and VMsNot documented
Machine sizeUp to 16 vCPU / 32 GB on the $100 plan and aboveNot published; billed per runtime-hour
Session lengthNo cap on paid plans; you set the time to liveHosted workspaces are built to run 24/7
State and snapshotsDisk saved every minute and on stop; a stopped sandbox costs nothingHosted workspace and synced browser profiles; no snapshot feature documented
ForkFork a running sandbox with its diskNot documented
BrowserChromium inside the VMLocal Chrome or Edge, or Opensteer Cloud browsers, driven over CDP with named sessions
Desktop and computer useA full Linux desktop streamed at 60fps; embed it, record itA billed computer runtime; what it contains is not published
AgentsSeven coding agents preinstalled, with a prompt, steer and events APIReady-made specialized agents on schedules; attach your local coding agent to a hosted workspace
Keeping API keys out of the sandboxCredentials proxy for model providers and any HTTPS API; switch a credential to another account when credits run outNo credentials proxy documented; its own API key is set as an environment variable
HostingHTTPS URLs for up to 50 ports, token-protected by default; raw TCPNo way to serve your own ports documented
Run it yourselfNoOpen-source MIT runtime that runs on your own machine with your own browser
Compliance and transparencySOC 2 is in progress. We're open about where we stand: live status and data retention policy.Not published
Every feature Battleships tracks (43)
FeatureboatOpensteer
Full VM (own kernel)YesNot published
Docker insideYesNo
Nested virtualizationYesNo
RootYesNot published
SystemdYesNot published
Desktop GUIYesYes
BrowserYesYes
Computer-use APIYesYes
SSHYesNo
Persistent diskYesYes
Pause/resumeYesNo
Automatic snapshotsYesNo
Fork/cloneYesNo
Live fork (no pause)YesNo
Memory forkNoNo
Custom image (Docker or snapshot)YesNot published
≥24 h sessionsYesYes
Idle auto-stopNoNo
Wake on requestNoNo
HTTPS preview URLsYesNo
Raw TCP inboundYesNot published
Public IPv4NoNo
Egress allowlistNoNot published
Secret proxyYesNo
Secret proxy for any APIYesNo
Preinstalled agentsYesYes
Agent harness APIYesYes
MCP serverNoNo
GPUNoNo
WindowsNoNo
Arm64NoNo
Self-hostingNoYes
BYOCNoNo
Open sourceNoYes
Scoped API keysYesNot published
Spend limitsYesNot published
Usage per sandbox (API)YesNo
Audit logsNoNot published
SSONoNot published
SOC 2NoNot published
HIPAANoNot published
EU data residencyYesNot published
Zero data retentionYesNot published

Performance

boat26.69 runs/s
OpensteerNot measured

There is no comparable measurement. Opensteer sells browser and computer runtime by the hour and does not publish the CPU or memory behind it, and it is not in the CPU benchmark. boat scored 26.69 runs per second on the hpc-sandbox-benchmarks Node.js test at 4 vCPU / 8 GB, on dedicated AMD Ryzen 9 9950X hosts.

Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Opensteer 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.

Scenarioboat, a monthOpensteer, 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$30 · Developer · with limits
One agent that never stops
4 vCPU / 8 GB, 50 GB disk, running 24/7 all month
$26 · $20/mo$141 · Developer · with limits
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$1,755 · Developer · with limits
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$995 · Developer · 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$1,995 · Developer · 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.

Opensteer: Opensteer bills $0.20 per hour of browser or computer runtime, paid from monthly platform credits ($5 free, $50 on Starter, $200 on Business). Model use is billed separately at API prices from model credits, unless you bring your own Claude or ChatGPT subscription.

Limits in those scenarios

None of the five scenarios fits Opensteer as written, because each asks for a general machine of a set size and Opensteer sells runtime-hours of a size it does not publish. It also documents no Docker, which the 50-agent scenario needs, and no limit on how many runtimes can run at once, which the 300 and 1,000 machine scenarios need. boat fits four scenarios; at 1,000 short jobs at once its $2,000 plan queues the starts.

ScenarioboatOpensteer
One coding agent, workdaysNone that bite
  • machine size not published
  • browser-session product, not a general sandbox
  • different kind of product: browser session
  • machine size per session not published
One agent that never stopsNone that bite
  • machine size not published
  • browser-session product, not a general sandbox
  • different kind of product: browser session
  • machine size per session not published
50 agents in parallelNone that bite
  • missing Docker inside
  • machine size not published
  • browser-session product, not a general sandbox
  • different kind of product: browser session
Sandboxes for your usersNone that bite
  • machine size not published
  • browser-session product, not a general sandbox
  • different kind of product: browser session
  • not confirmed: how many machines can run at once (you need 300; the typical published limit is 50)
1,000 short jobs at once
  • max 2,400 starts/day (needs ~6,667)
  • max 600 starts/hour (needs ~1,000)
  • bursts above 90 starts/min are queued (a burst of 333 takes ~4 min to start)
  • machine size not published
  • browser-session product, not a general sandbox
  • different kind of product: browser session
  • not confirmed: how many machines can run at once (you need 1,000; the typical published limit is 50)

Which one should you use?

Use Opensteer when

you want ready-made browser agents on a schedule, or an open-source runtime that lets your local coding agent drive your own logged-in Chrome.

  • Open source, on your own machine. For agents that should use your own logged-in Chrome or Edge: the MIT runtime runs locally and drives it over CDP.
  • Ready-made agents with schedules. For teams that want finished browser agents: they run on schedules, with synced browser profiles and team access.
  • Use the AI plan you already pay for. For people with a Claude or ChatGPT subscription: agents can run on it instead of paid model credits.

Use boat when

your agent does more than browse: it needs a shell, Docker, a desktop, files that are saved and forked, or a server next to the browser, on a machine of known size.

  • A whole machine around the browser. Opensteer's hosted workspace is reached through a command-and-file toolbelt, with no SSH, no documented Docker and no published machine size. Work that needs containers or a server has to run somewhere else.
  • State you can save, stop and fork. Opensteer documents no snapshots, pause or fork, so its workspaces keep running to keep their state. Starting ten agents from one prepared machine means setting up each one again.
  • Any coding agent, driven from outside, with keys kept out. Opensteer documents no credentials proxy, so keys for the APIs your agent calls sit where the agent can read them. One prompt injection in a page it browses can then leak them.

Questions people ask

Is Opensteer a sandbox like boat?

No. Opensteer is an agent product: hosted agents with browser and computer runtime, reached through its own command and file tools rather than SSH. boat gives you a whole Linux VM with root, Docker, Chromium and a desktop, so the browser is one tool among many.

Is boat cheaper than Opensteer?

In four of the five scenarios on this page, yes, for example $26 against $141 for one agent running all month. For 1,000 short jobs at once the two are about the same, and neither fits. Opensteer's model use is billed on top of runtime.

Can I run Opensteer's open-source runtime on boat?

It is MIT licensed and installs with one command, and a boat sandbox has Python, Chromium and root, so you can install it there. It is built for local Chrome or Edge, so check that it attaches to the browser you run in the sandbox.

Can Opensteer run Docker or an Android emulator?

Opensteer documents neither; its workspaces are for browser and computer-use agents. A boat sandbox runs Docker and systemd and exposes /dev/kvm for Android emulators and VMs.

How big is an Opensteer machine?

Opensteer does not publish the vCPU or memory behind its browser and computer runtime. A boat sandbox defaults to 4 vCPU / 8 GB and goes up to 16 vCPU / 32 GB on the $100 plan and above.

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

More comparisons