Steel vs Browserless
Both rent you cloud browsers that your agent drives over CDP. Steel runs each headful Chrome session in its own VM in us-east and bills by the minute. Browserless serves Chrome, Firefox and WebKit from a managed pool in the US and Europe and bills in 30-second units.
The companies and products
Steel
Steel is an API that runs fleets of cloud browsers for AI agents: you create a session, connect Puppeteer, Playwright or Selenium, and release it when done.
Each Steel session is a headful Chromium in a VM in Steel's cloud. You drive it over CDP, a Commands API or a computer-use input API, and you can embed a live viewer. Around it Steel sells managed proxies, CAPTCHA solving, saved browser profiles, a credentials vault that signs in for the agent, a files API and readable traces. It is a different kind of product from boat: a managed browser, not a general sandbox. This page compares where the two overlap: agents that use a browser.
Browserless
Browserless calls itself a cloud browser automation platform: managed browsers you connect to with one line of code.
You point existing Puppeteer, Playwright or Selenium code at a Browserless WebSocket URL. You can also use BrowserQL, its GraphQL language with stealth and CAPTCHA solving, REST endpoints for screenshots, PDFs and scraping, an MCP server, a CLI and agent skills. Self-serve sessions last 2 to 60 minutes by plan. Between sessions, the browser profile can be kept, and processKeepAlive (Puppeteer) keeps the live browser for a set window. Enterprise adds persistent sessions on dedicated workers. It is a browser product, not a general sandbox: your code gets a browser, not a shell on the machine. This page compares where the two overlap, which is agents that browse.
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. “24th cheapest of 356” means 23 of the other 355 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 | Steel: price rank | Browserless: price rank | boat: price rank |
|---|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A coding agent needs a shell, Docker and a disk. Neither hosted service gives one, so installs, builds and tests need a machine like boat's. | 157th of 365, with limits $1,115 a month | 198th of 365, with limits $1,771 a month | 5th cheapest of 365 $634 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB An assistant that works all month hits a 1-hour session cap on both Scale plans and reloads its browser profile each time; Browserless Enterprise adds persistent sessions on dedicated workers. A boat sandbox keeps running once auto-stop is off (default 1 hour). | 128th of 361, with limits $3,070 a month | 311th of 361, with limits $6,170 a month | 6th cheapest of 361 $657 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month | 262nd of 336, with limits $333 a month | 265th of 336, with limits $350 a month | 189th of 336, with limits $2,000 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month 300 sandboxes at once is more than the 100 browsers both Scale plans allow, so both need Enterprise, and each user gets a browser rather than a machine. | 240th of 349, with limits $990 a month | 258th of 349, with limits $1,490 a month | 3rd cheapest of 349 $209 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 252nd of 349, with limits $4,150 a month | 273rd of 349, with limits $8,600 a month | 39th of 349, with limits $2,000 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month | 345th of 359, with limits $400 a month | 251st of 359, with limits $350 a month | 10th cheapest of 359 $200 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month | 239th of 374, with limits $120 a month | 242nd of 374, with limits $140 a month | 2nd cheapest of 374 $86 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB A dev box has to keep its files and processes. Both products keep browser state between sessions, not a disk, so this fits boat. | 255th of 371, with limits $73 a month | 341st of 371, with limits $350 a month | 19th cheapest of 371 $26 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Steel and Browserless let an agent act on web pages, with a person able to take over; boat streams a 60fps Linux desktop where the agent can use any app. | 190th of 367, with limits $367 a month | 183rd of 367, with limits $350 a month | 7th cheapest of 367 $200 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month All three fit: our default ranking puts boat at #6, Browserless at #13 and Steel at #18. boat ranks well when the same agent also needs a shell and Docker next to its Chromium. | 24th cheapest of 356 $470 a month | 19th cheapest of 356 $350 a month | 7th cheapest of 356 $200 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month All three fit at startup scale. Steel adds dedicated IPs on Scale, Browserless adds REST endpoints for screenshots and PDFs, and boat runs the scraper and its database on one VM. | 61st cheapest of 343 $540 a month | 50th cheapest of 343 $478 a month | 6th cheapest of 343 $90 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month For sites that block bots, our default ranking puts Steel at #5 and Browserless at #11, both with residential proxies and CAPTCHA solving, and boat at #46 on datacenter IPs. | 8th of 327, with limits $7,950 a month | 19th of 327, with limits $17,150 a month | 73rd of 327, with limits $1,800 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 250th of 351, with limits $242 a month | 113th of 351, with limits $350 a month | 3rd cheapest of 351 $200 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month | Not ranked: misses a requirement | Not ranked: misses a requirement | Not ranked: misses a requirement |
| macOS agents 5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month | Not ranked: misses a requirement | Not ranked: misses a requirement | Not ranked: misses a requirement |
Key features
| Feature | Steel | Browserless | boat |
|---|---|---|---|
| Kind of product | A managed headful Chromium per session, driven over CDP, REST, a Commands API or a computer-use input API | Managed Chrome, Firefox and WebKit, plus BrowserQL, REST APIs, an MCP server, a CLI and agent skills | A full Linux VM per sandbox, with a desktop and Chromium inside |
| Isolation | Each session is a Chromium in a VM, per the docs; VM type not named | Managed browser pool with resource isolation between sessions; isolation technology not published | Full VM with its own kernel, on physical hosts dedicated to boat (AMD Ryzen 7 9700X and Ryzen 9 9950X); vCPUs are shared |
| Browsers | Chromium, headful | Chrome, Firefox and WebKit | Chromium on a 60fps desktop |
| Machine size | Not published; billed per browser hour | Not published; billed in 30-second units | 4 vCPU / 8 GB by default, 2 to 16 vCPU; 16 vCPU / 32 GB needs the $100 plan or higher |
| Shell, Docker and nested virtualization | A hosted session gives browser control only; you can self-host the open-source steel-browser in Docker next to your own code | A hosted session runs a browser, with no shell or SSH; you can run the Browserless Docker image on your own host | SSH, root, systemd, Docker and /dev/kvm on every machine |
| Session length | 15 minutes on Launch, 1 hour on Scale, up to 24 hours on Enterprise; longer jobs chain sessions | 2 minutes on Free, 15 on Prototyping, 30 on Starter, 60 on Scale; Enterprise persistent sessions on dedicated workers can stay running | Auto-stop defaults to 1 hour; set up to 30 days, or turn it off on a paid plan (trial: 2 hours maximum) |
| What stays after a session | Profiles keep cookies and logins (300 MB, deleted after 30 days unused); session files become org-wide files that mount into later sessions | Persisted sessions keep cookies, localStorage and cache for 1 to 90 days by plan; by default a new browser process starts on reconnect, and processKeepAlive (Puppeteer only) keeps the live browser for a set window | The disk, saved every minute and at each stop, kept at least 14 days after a stop; memory and /tmp are not saved; fork a running sandbox into a copy |
| Computer use and human takeover | Computer-use input API, embeddable live view with human takeover, mobile mode and traces | Live URLs that hand a session to a person, and Anthropic and OpenAI computer-use integrations | 60fps desktop, computer use on any Linux app, embeddable stream |
| Bot detection | Residential proxies, CAPTCHA solving and dedicated IPs; a stealth browser on Enterprise | BrowserQL stealth with fingerprint mitigations, CAPTCHA solving, residential and datacenter proxies | Datacenter IPs and Chromium; no built-in CAPTCHA solving; you set up proxies or a CAPTCHA service yourself |
| Reaching the machine | No inbound ports; a CDP websocket and a live view URL | Fixed regional WebSocket and HTTPS endpoints and short-lived live URLs; no ports to host on | Up to 50 HTTPS URLs per sandbox, token-gated by default; SSH; a public IPv6 address; no dedicated IPv4 |
| Agents | A CLI, skills and an MCP server give your coding agent a Steel browser; the agent runs on a machine you provide | Its own managed browser agent (Agent Run API, beta), up to 15 minutes and 60 steps per turn, plus agent skills and a Claude Code plugin | Seven coding-agent harnesses installed (Claude Code, Codex, OpenCode and more), with a prompt, steer and events API |
| Secrets | Credentials API (beta) stores web logins and signs in without showing the agent the password | 1Password integration fills logins into the page without the value reaching your code | Optional paid AI credentials proxy ($1 a day it is on, plus $1 per 1,000 requests): real keys of supported model providers stay outside the sandbox; other secrets still enter it |
| Self-hosting and license | Open-source steel-browser (Apache-2.0) in Docker, and Steel Outpost inside your own network | Open source core under SSPL-1.0; commercial self-hosting needs a paid license | Managed cloud only |
| Compliance | HIPAA-ready BAA on Scale and Enterprise; SSO on Enterprise | SOC 2 Type II, GDPR, HIPAA with a BAA, and a DPA, listed on the Enterprise plan | SOC 2 in progress; no HIPAA BAA or SAML SSO documented |
| Regions | us-east only; proxy location sets the visible IP; Steel Outpost runs in your own region | San Francisco, London and Amsterdam; New York on private fleets | Germany, Finland and France; no US or Asia region; you cannot pick one |
Every feature Battleships tracks (43)
| Feature | Steel | Browserless | boat |
|---|---|---|---|
| Full VM (own kernel) | Yes | Not published | Yes |
| Docker inside | No documented way | No documented way | Yes |
| Nested virtualization | No documented way | No documented way | Yes |
| Root | No documented way | Not published | Yes |
| Systemd | No documented way | Not published | Yes |
| Desktop GUI | No documented way | No documented way | Yes |
| Browser | Yes | Yes | Yes |
| Computer-use API | Yes | Yes | Yes |
| SSH | No documented way | No documented way | Yes |
| Persistent disk | Yes | No documented way | Yes |
| Pause/resume | No documented way | No documented way | Yes |
| Automatic snapshots | Yes | No documented way | Yes |
| Fork/clone | No documented way | No documented way | Yes |
| Live fork (no pause) | No documented way | No documented way | Yes |
| Memory fork | No documented way | No documented way | No documented way |
| Custom image (Docker or snapshot) | No documented way | Not published | Yes |
| ≥24 h sessions | Yes | No documented way | Yes |
| Idle auto-stop | Yes | Yes | No documented way |
| Wake on request | No documented way | No documented way | No documented way |
| HTTPS preview URLs | No documented way | No documented way | Yes |
| Raw TCP inbound | No documented way | Not published | Yes |
| Public IPv4 | No documented way | No documented way | No documented way |
| Egress allowlist | No documented way | Yes | No documented way |
| Secret proxy | No documented way | No documented way | Yes |
| Secret proxy for any API | No documented way | No documented way | Yes |
| Preinstalled agents | No documented way | Yes | Yes |
| Agent harness API | No documented way | No documented way | Yes |
| MCP server | Yes | Yes | No documented way |
| GPU | No documented way | Yes | No documented way |
| Windows | No documented way | No documented way | No documented way |
| Arm64 | No documented way | No documented way | No documented way |
| Self-hosting | Yes | Yes | No documented way |
| BYOC | Yes | Yes | No documented way |
| Open source | Yes | Yes | No documented way |
| Scoped API keys | Not published | Not published | Yes |
| Spend limits | Not published | Not published | Yes |
| Usage per sandbox (API) | No documented way | No documented way | Yes |
| Audit logs | Not published | Not published | No documented way |
| SSO | Yes | Yes | No documented way |
| SOC 2 | No documented way | Yes | No documented way |
| HIPAA | Yes | Yes | No documented way |
| EU data residency | No documented way | Yes | No documented way |
| Zero data retention | Not published | Not published | Yes |
Performance
Steel and Browserless are not in the public benchmarks we use, and neither publishes the vCPU or RAM behind a browser session, so we cannot say which is faster. For reference, a boat sandbox at 4 vCPU / 8 GB did 28.06 runs/s on the V8 Web Tooling Benchmark in StarSling's run of 3 October 2026 (run 37073680030), second of 15. Every boat sandbox in this run ran on an AMD Ryzen 9 9950X host. boat also places some sandboxes on AMD Ryzen 7 9700X hosts, which this run did not measure. Steel says a session starts in under a second.
V8 Web Tooling Benchmark at a 4 vCPU / 8 GB target, runs per second, higher is better. Source: StarSling's run of 3 October 2026 (run 37073680030), median of 5 sandboxes. Providers ran on different hardware in different regions. StarSling gives a shared rank when the run cannot separate two providers. Every boat sandbox in this run ran on an AMD Ryzen 9 9950X host. boat also places some sandboxes on AMD Ryzen 7 9700X hosts, which this run did not measure. Steel and Browserless are not in the public benchmarks we use, so we cannot say which is faster.
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. Start limits, per-start fees and plan minimums are in the price, boat's included.
| Scenario | Steel, a month | Browserless, a month | boat, 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 | $250 · Scale · with limits | $350 · Scale annual · with limits | $20 · $20/mo |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $250 · Scale · with limits | $350 · Scale annual · with limits | $26 · $20/mo |
| 50 agents in parallel 50 machines of 4 vCPU / 8 GB at once, 8-hour sessions, 1,100 sessions a month, Docker inside | $784 · Scale · with limits | $1,026 · Scale annual · with limits | $317 · $20/mo × 3 seats |
| Sandboxes for your users 2 vCPU / 4 GB, 30,000 sessions of 10 minutes a month, up to 300 at once | $510 · Scale · with limits | $410 · Scale annual · with limits | $200 · $100/mo × 2 seats |
| 1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month | $950 · Scale · with limits | $1,400 · Scale annual · with limits | $8,800 · $2000/mo · with limits |
Steel: Steel bills browser hours by the minute, rounded up. Launch has no monthly fee, charges $0.10 an hour and gives a one-time $30 credit valid for 90 days. Scale costs $250 a month, includes $100 of usage and charges $0.08 an hour. Proxy bandwidth, CAPTCHA solves and browser tools are billed on top.
Browserless: Browserless bills in units: one unit is up to 30 seconds of a connected browser, and a reconnect starts a new block. Annual plans cost $25, $140 or $350 a month for 20,000, 180,000 or 500,000 units, then $0.0020, $0.0017 or $0.0015 a unit. Proxy traffic uses 6 or 2 units per MB and a CAPTCHA solve uses 10 units. There is a $0 Free plan with 2-minute sessions.
Browserless: Plans cost $20, $100, $500 or $2,000 a month, and the whole plan price comes back as VM time for that month. A default 4 vCPU / 8 GB sandbox costs $0.036 an hour, billed per second while it runs, busy or idle. A stopped sandbox has no compute charge and keeps its disk for at least 14 days; saved snapshots above the included count and the optional AI credentials proxy still charge. The $20 plan includes 12 starts a minute and 200 a day; extra starts cost $0.03 to $0.07. A card is required, also for the 7-day trial.
Limits in those scenarios
Neither Steel nor Browserless fits any scenario in the table, because each sells a browser session with no published size, no shell and no disk that keeps installed state. On Steel, work that runs for hours must be cut into sessions: 15 minutes on Launch, 1 hour on Scale, 24 hours on Enterprise. Browserless Scale stops a session after 60 minutes, and Enterprise persistent sessions are priced by sales. Both cap you at 100 browsers at once on Scale, so 300 sandboxes for your users and 1,000 short jobs at once need Enterprise. On boat, the 1,000-job burst does not fit either: even the $2,000 plan includes 90 starts a minute and 2,400 a day, extra starts cost $0.03 to $0.07, starts above 270 a minute fail with 429, and boat never queues a start. A boat sandbox also stops after 1 hour by default unless you turn auto-stop off on a paid plan.
| Scenario | Steel | Browserless | boat |
|---|---|---|---|
| One coding agent, workdays |
|
| None for this scenario |
| One agent that never stops |
|
| None for this scenario |
| 50 agents in parallel |
|
| None for this scenario |
| Sandboxes for your users |
|
| None for this scenario |
| 1,000 short jobs at once |
|
|
|
Which one should you use?
Use Steel when
you want each browser in its own VM, an Apache-2.0 runtime you can host yourself, no monthly fee to start, sessions up to 24 hours on Enterprise, or ready-made logins, takeover and traces in the US.
- A VM behind every browser. Useful if you want each session kept apart: Steel's docs describe every headful Chromium as running in its own VM.
- Permissive open source, and self-hosted. Useful if you will run browsers yourself: steel-browser is Apache-2.0, and Steel Outpost runs inside your own network.
- Longer sessions and agent tooling. Useful for long browser runs: Enterprise sessions last up to 24 hours, with a login vault, computer-use API and traces.
Use Browserless when
you need Firefox or WebKit as well as Chrome, browsers in Europe or San Francisco, one-call REST screenshots and PDFs, or BrowserQL stealth for sites that block bots.
- Three browser engines. Useful if you test or scrape beyond Chrome: Browserless serves Chrome, Firefox and WebKit from the same API.
- Browsers in Europe as well as the US. Useful if your traffic should start near Europe: the shared fleet runs in London and Amsterdam as well as San Francisco.
- Jobs with no script. Useful for one-call tasks: REST endpoints return screenshots, PDFs and scraped content, and BrowserQL adds stealth and CAPTCHA solving.
Where boat fits
Steel and Browserless are both browsers, not general sandboxes. If what you are really running is agents that browse and then also install, build, run Docker or keep files, look at boat: each agent gets a full Linux VM with Chromium on a desktop inside it. If the job is only scraping sites that block bots, Steel or Browserless will serve you better.
A whole machine, not only a browser
Neither hosted service gives a shell, Docker or a published machine size. An agent that reads a page and then clones a repo or starts a dev server needs a second machine, and you move files between the two.
Long runs and a disk that stays
Both Scale plans stop a session after 1 hour, and a browser profile carries over; longer sessions need Enterprise. On boat a paid sandbox runs as long as the task once auto-stop is off, saves its disk every minute, and can be forked while running.
Coding agents built in
With Steel you host the agent yourself, and Browserless's beta agent stops at 15 minutes and 60 steps a turn. boat ships seven coding-agent harnesses, including Claude Code, Codex and OpenCode, that you drive from outside.
Use boat when your agent browses and also works: installs, builds, Docker and dev servers on a full VM, a disk that is saved every minute, a 60fps desktop and coding agents built in. Compare it one to one: boat vs Steel and boat vs Browserless.
boat's own limits
- Plans start at $20 a month. The plan price comes back as VM time for that month. Unused time does not carry over. Source
- Starts (create, fork, resume): the $20 plan includes 12 a minute, 60 an hour and 200 a day. The $2,000 plan includes 90 a minute, 600 an hour and 2,400 a day. Extra starts cost $0.03 to $0.07. A start fails with 429 above 3 times the included starts per minute, or 6 times per hour or per day. Source
- A stopped sandbox has no compute charge. boat keeps its disk for at least 14 days. A saved snapshot keeps it longer; extra saved snapshots and the AI credentials proxy are paid. Source
- Auto-stop defaults to 1 hour. You can turn it off on a paid plan. The trial allows 2 hours at most. Source
- boat saves the disk every minute and at each stop. Memory, running programs and /tmp are not saved: a resume is like a reboot. A machine failure loses writes after the last good snapshot. Source
- Machines run in Germany, Finland and France only. Snapshot copies can be stored outside the EU. Source
- Public IPv6 and outbound IPv4. No dedicated IPv4. The IP changes at each resume. Source
- Hosts are AMD Ryzen 7 9700X and Ryzen 9 9950X. vCPUs are shared. No GPU. You cannot boot your own Docker image. Source
Questions people ask
Is Steel cheaper than Browserless?
For pure browser work, Browserless usually is. In our default rankings, browser-use agents cost $350 a month on Browserless and $470 on Steel, and startup-scale scraping $478 against $540. boat comes to $200 and $90. One agent on workdays costs about $18 on Steel's Launch plan if it restarts every 15 minutes from a saved profile, or $250 on Scale for 1-hour sessions; on Browserless it does not fit, because Scale ($350) ends sessions after 60 minutes and keeps no disk. For 300 browsers or 1,000 short jobs at once, both need a custom Enterprise quote.
Which is faster, Steel or Browserless?
Neither is in the public benchmarks we use, and neither publishes the vCPU or RAM behind a session, so we cannot say. Steel says a session starts in under a second.
Can Steel or Browserless run Docker?
Not in a hosted session: on either, it is a browser you reach over CDP and an API, with no shell. Both let you run their browser image in Docker on your own servers. Every boat sandbox has root, SSH, Docker, systemd and /dev/kvm.
How long can a session run on Steel and Browserless?
Steel allows 15 minutes on Launch, 1 hour on Scale and 24 hours on Enterprise. Browserless allows 2 minutes on Free up to 60 minutes on Scale, and Enterprise persistent sessions on dedicated workers can stay running. A boat sandbox stops after 1 hour by default, and you can turn auto-stop off on a paid plan.
Which one is open source?
Both publish their core. Steel's steel-browser is under Apache-2.0, so you can run it freely. Browserless is under SSPL-1.0 and free for non-commercial use, and commercial use needs a paid license.
Can I run browsers in Europe?
Browserless's shared fleet runs in London and Amsterdam as well as San Francisco. Steel's hosted browsers run in us-east, and proxies change only the visible IP; Steel Outpost runs in your own network. boat runs on machines in Germany, Finland and France, with no US region.
Try boat
Install the CLI, then run boat new.
7-day free trial