Steel vs Anchor Browser
Both rent your agent a cloud Chromium it drives over CDP. Steel is open source and you can host it yourself, but its sessions stop after 15 minutes to 24 hours by plan; Anchor Browser is not open source, lets sessions run as long as you set, and handles logins with MFA, stealth and more regions for you.
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, driven over CDP or through a computer-use input API, with a live viewer you can embed. 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 bochka: a managed browser, not a general sandbox, so this page compares where the two overlap, which is agents that use a browser.
Anchor Browser
Anchor Browser is cloud browser infrastructure for AI agents: managed Chromium sessions you control over CDP or a REST API, plus a hosted browser agent.
Anchor Browser is a different kind of product from bochka. Each session is a headful Chromium in its own VM that is erased when the session ends, and you drive it with Playwright or Puppeteer, with mouse and keyboard calls, or by sending a task in plain words to its hosted agent. It adds managed logins with MFA, residential proxies, stealth and CAPTCHA solving, and its own docs tell you to run your agent's code in a separate sandbox and connect to the browser over CDP.
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. “25th cheapest of 367” means 24 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 | Steel: price rank | Anchor Browser: price rank | bochka: price rank |
|---|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month Neither has a shell, Docker or a place to build and test; Anchor's own docs send that code to a separate sandbox, and Steel's coding-agent tools run the agent on your machine. | 124th of 367, with limits $1,115 a month | 180th of 367, with limits $2,000 a month | 3rd cheapest of 367 $634 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB An assistant that works all month suits Anchor Browser better than Steel, whose sessions stop after 1 hour on Scale, but both erase the machine and keep only profiles and files. | 238th of 366, with limits $3,070 a month | 71st of 366, with limits $1,845 a month | 5th cheapest of 366 $657 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month | 264th of 345, with limits $333 a month | 325th of 345, with limits $2,171 a month | 186th of 345, 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 Steel's Scale limit of 100 and Anchor's Growth limit of 200, and each user gets a browser rather than a machine. | 244th of 361, with limits $990 a month | 283rd of 361, with limits $2,000 a month | Cheapest of 361 $209 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 257th of 360, with limits $4,150 a month | 298th of 360, with limits $12,510 a month | 34th of 360, with limits $2,000 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month | 348th of 364, with limits $400 a month | 357th of 364, with limits $2,000 a month | 10th cheapest of 364 $200 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month | 239th of 377, with limits $120 a month | 284th of 377, with limits $500 a month | 4th cheapest of 377 $86 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB A dev box needs SSH and a shell, and neither Steel nor Anchor Browser exposes more than a CDP socket and a live view. | 241st of 368, with limits $73 a month | 222nd of 368, with limits $42 a month | 18th cheapest of 368 $26 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Steel and Anchor Browser both give a computer-use input API inside the browser; bochka streams a full Linux desktop at 60fps, so the agent can use apps beyond the browser. | 148th of 369, with limits $367 a month | 173rd of 369, with limits $500 a month | 6th cheapest of 369 $200 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month This is what both are built for and they sit close: Steel ranks 25th and Anchor Browser ranks 28th, while bochka ranks 7th because the same machine can also run the agent's code. | 25th cheapest of 367 $470 a month | 28th cheapest of 367 $560 a month | 7th cheapest of 367 $200 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month Both fit startup-scale bots with proxies and CAPTCHA solving; bochka fits too and keeps the scraper, its queue and its database on one machine whose disk stays. | 56th cheapest of 346 $540 a month | 50th cheapest of 346 $500 a month | 6th cheapest of 346 $90 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month Strong bot checks favor the two browser services: Anchor Browser ranks 5th with stealth Chromium on Growth, Steel ranks 9th with its stealth browser only on Enterprise, and bochka uses datacenter IPs. | 9th of 338, with limits $7,950 a month | 5th of 338, with limits $6,175 a month | 71st of 338, with limits $1,800 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 220th of 360, with limits $242 a month | 334th of 360, with limits $2,000 a month | 2nd cheapest of 360 $200 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 | 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 | Can't run it |
Key features
| Feature | Steel | Anchor Browser | bochka |
|---|---|---|---|
| Kind of product | A managed headful Chromium per session, driven over CDP or the REST API | A managed headful Chromium per session, over CDP or REST, plus hosted browser agents | A full Linux VM per sandbox, with a desktop and Chromium inside |
| Isolation | Each session is a Chromium in a VM, per the docs; the VM type is not named | A dedicated VM per session, terminated and erased when the session ends | Every sandbox is a full VM with its own kernel, on dedicated AMD Ryzen 9 9950X hosts |
| Machine size | Not published; billed per browser hour | Not published; billed per browser hour | 4 vCPU / 8 GB by default, up to 16 vCPU / 32 GB by plan |
| Shell, code and Docker | No shell; you reach the browser through CDP and the API | No shell; from Growth you can run TypeScript snippets against the page | SSH, root, any language, Docker, systemd and /dev/kvm on every machine |
| Session length | 15 minutes on Launch, 1 hour on Scale, up to 24 hours on Enterprise | 180 minutes by default, with no upper limit if you raise it; ends after 5 idle minutes by default | No session cap on paid plans; a stopped sandbox keeps its disk |
| State after a session | Files API keeps session files; profiles keep cookies and logins (300 MB, deleted after 30 days unused) | The VM is erased; profiles keep cookies, local storage and cache, and up to 10 volumes keep files | The whole disk is saved every minute and on stop; a stopped sandbox costs nothing |
| Fork | No fork; a new session can start from a saved profile or session context | No fork; a new session can start from a saved profile | Fork a running sandbox into a copy |
| Bot checks and proxies | Residential proxies, CAPTCHA solving on paid plans, dedicated IPs; stealth browser on Enterprise | CAPTCHA solving from Starter, stealth Chromium on Growth, unique fingerprints on Enterprise; city-level residential proxies | Chromium on datacenter IPs; no CAPTCHA solving |
| Logging in to sites | Profiles, plus a Credentials API (beta) that signs in without showing the agent the password | Managed identities with MFA keep logins across sessions, from Starter | Your agent logs in itself; the credentials proxy keeps real API keys outside the sandbox |
| Agents | A CLI, skills and an MCP server give your coding agent a Steel browser; the agent runs elsewhere | Hosted browser agents (browser-use, OpenAI, Gemini, Anthropic computer use) with human-in-the-loop and webhooks | Claude Code, Codex, OpenCode and more preinstalled, driven from outside with a prompt, steer and events API |
| Live view and computer use | Embeddable live view with human takeover, a computer-use input API, mobile mode and traces | Embeddable live view of the browser, with mouse, keyboard and screenshot calls | A full Linux desktop streamed at 60fps, or Chromium alone, embeddable in your app |
| Reaching the machine | No inbound ports; a CDP websocket and a live view URL | No ports of your own; only the CDP socket and the live view URL | HTTPS URLs for up to 50 ports, token-protected; raw TCP |
| Self-hosting | Open-source steel-browser (Apache-2.0) in Docker, and Steel Outpost inside your own network | Your own cloud or on-premise on Enterprise | Managed cloud only |
| Regions | us-east only; proxy location changes only the visible IP | US by default; Frankfurt, Asia and Australia on Growth; any region on Enterprise | EU only: Germany, Finland, France |
Every feature Battleships tracks (43)
| Feature | Steel | Anchor Browser | bochka |
|---|---|---|---|
| Full VM (own kernel) | Yes | Yes | Yes |
| Docker inside | No | No | Yes |
| Nested virtualization | No | No | Yes |
| Root | No | No | Yes |
| Systemd | No | No | Yes |
| Desktop GUI | No | No | Yes |
| Browser | Yes | Yes | Yes |
| Computer-use API | Yes | Yes | Yes |
| SSH | No | No | Yes |
| Persistent disk | No | Yes | Yes |
| Pause/resume | No | No | Yes |
| Automatic snapshots | Yes | No | Yes |
| Fork/clone | No | No | Yes |
| Live fork (no pause) | No | No | Yes |
| Memory fork | No | No | No |
| Custom image (Docker or snapshot) | No | No | Yes |
| ≥24 h sessions | Yes | Yes | Yes |
| Idle auto-stop | Yes | Yes | No |
| Wake on request | No | No | No |
| HTTPS preview URLs | No | No | Yes |
| Raw TCP inbound | No | No | Yes |
| Public IPv4 | No | No | No |
| Egress allowlist | No | Yes | No |
| Secret proxy | No | No | Yes |
| Secret proxy for any API | No | No | Yes |
| Preinstalled agents | No | Yes | Yes |
| Agent harness API | No | No | Yes |
| MCP server | Yes | Yes | No |
| GPU | No | No | No |
| Windows | No | No | No |
| Arm64 | No | No | No |
| Self-hosting | Yes | Yes | No |
| BYOC | Yes | Yes | No |
| Open source | Yes | No | No |
| Scoped API keys | Not published | Not published | Yes |
| Spend limits | Not published | Not published | Yes |
| Usage per sandbox (API) | No | Yes | Yes |
| Audit logs | Not published | Not published | No |
| SSO | Yes | Yes | No |
| SOC 2 | No | Yes | No |
| HIPAA | Yes | Yes | No |
| EU data residency | No | Yes | Yes |
| Zero data retention | Not published | Not published | Yes |
Performance
There is no comparable measurement for either. Neither Steel nor Anchor Browser publishes the CPU or memory behind a browser session, and neither is in the StarSling CPU benchmark. bochka scored 26.69 runs per second on its Node.js test at 4 vCPU / 8 GB; Steel says its sessions start in under a second.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Steel and Anchor Browser have 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 | Steel, a month | Anchor Browser, a month | bochka, 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 | $7 · Free · 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 | $42 · Free · 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 | $500 · Team · 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 | $2,000 · Growth · 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 | $2,510 · Growth · with limits | $2,000 · $2000/mo · with limits |
Steel: Steel bills browser hours by the minute, rounded up: $0.10 an hour on the free Launch plan and $0.08 on Scale, which costs $250 a month and includes $100 of usage. Proxy bandwidth and CAPTCHA solves are billed on top.
Anchor Browser: Anchor Browser bills $0.05 per browser-hour plus $0.01 for each browser created, paid from credits; plan fees (Starter $50, Team $500, Growth $2,000) turn into credits. Proxy traffic is $8 per GB, direct traffic $0.20 per GB, and AI tasks are priced per step.
Anchor Browser: Plans start at $20 a month and the whole fee is usage credit, spent per second while a sandbox runs. A stopped sandbox costs nothing and keeps its disk.
Limits in those scenarios
None of the five scenarios fits Steel or Anchor Browser as written, because each asks for a general machine and both sell a browser session with no shell and no published size. Steel needs Scale for anything past 15 minutes and still stops each session after 1 hour, while Anchor runs 180 minutes by default and erases the VM each time. For 300 sandboxes at once, Steel's Scale cap of 100 and Anchor's Growth cap of 200 both fall short. bochka fits four scenarios; at 1,000 short jobs at once its $2,000 plan queues the starts.
| Scenario | Steel | Anchor Browser | bochka |
|---|---|---|---|
| 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 Steel when
you want an open-source browser you can host yourself, or many short browser sessions with a live view, human takeover and traces, in the US.
- Open source you can host. For teams that want to read the code or run browsers on their own machines: steel-browser is Apache-2.0 and runs in Docker.
- More browsers on a smaller plan. For fleets of short browser jobs: Steel's Scale plan allows 100 browsers at once, where Anchor needs Growth for more than 50.
- Ready-made browser tooling. For agents that need a person nearby: a live view with human takeover, mobile mode, readable traces and a credentials vault.
Use Anchor Browser when
sessions run for hours and the hard part is the website: logins with MFA, stealth, CAPTCHAs and proxies, in the US, EU or Asia.
- Long sessions. For agents that work for hours: sessions run 180 minutes by default and you can raise that with no upper limit.
- Logins and stealth handled. For hard websites: managed logins with MFA from Starter, CAPTCHA solving, and stealth Chromium on Growth rather than Enterprise.
- More regions and hosted agents. For teams outside the US or that want to send a task in plain words: EU, Asia and Australia on Growth, plus hosted agents.
Where bochka fits
Steel and Anchor Browser both sell a browser, not a machine. If what you are really running is agents that browse and then install, build, test or serve what they found, look at bochka: a full Linux VM with Chromium and a desktop inside, where the browser is one tool among many.
A whole machine around the browser
Neither Steel nor Anchor Browser gives a shell or Docker. Anchor's docs tell you to run agent code in a separate sandbox, and Steel's tools run the agent on your machine, so you pay for and wire up two systems.
Work that is kept between runs
Steel stops a session after 1 hour on Scale and Anchor erases the VM when a session ends; both carry over only profiles and files. On bochka the whole disk is saved every minute, stops are free, and you can fork a running sandbox.
A desktop, ports and keys outside
Both stop at the browser and open no ports. bochka streams a 60fps Linux desktop for any app, serves HTTPS on up to 50 ports and raw TCP, and keeps real API keys outside the sandbox.
Use bochka when your agent does more than browse: it needs a shell, Docker, a desktop or a server next to the browser, and has to keep its work between runs. Compare it one to one: bochka vs Steel and bochka vs Anchor Browser.
Questions people ask
Is Steel cheaper than Anchor Browser?
It depends on the plan you need. Anchor's rate is $0.05 per browser-hour against Steel's $0.08 to $0.10, and one coding agent on workdays comes to $7 on Anchor against $250 on Steel, since Steel needs Scale for sessions over 15 minutes. For 300 sandboxes for your users it flips: $510 on Steel against $2,000 on Anchor, and neither fits.
Which is faster, Steel or Anchor Browser?
There is no comparable measurement. Neither publishes the CPU or memory behind a session, and neither is in the StarSling CPU benchmark. Steel says its sessions start in under a second.
Can Steel or Anchor Browser run Docker?
No. Both give you a browser you drive over CDP and an API, with no shell on the machine. Anchor's docs say to run your agent's code in a separate sandbox; bochka gives each sandbox a full VM with Docker and root.
How long can a session run on Steel and on Anchor Browser?
Steel sessions stop after 15 minutes on Launch, 1 hour on Scale and 24 hours on Enterprise, and a live session cannot be extended. Anchor sessions run 180 minutes by default, and you can raise that with no upper limit and turn off the 5-minute idle timeout.
Which is better at getting past bot checks?
Both are built for it, with residential proxies and CAPTCHA solving. Anchor puts stealth Chromium on its Growth plan and unique fingerprints on Enterprise, while Steel's stealth browser is Enterprise only. bochka runs Chromium on datacenter IPs and solves no CAPTCHAs.
Can I self-host Steel or Anchor Browser?
Steel's steel-browser runtime is open source under Apache-2.0 and runs in Docker, and Steel Outpost runs inside your own network. Anchor offers your own cloud or on-premise only on Enterprise. bochka is a managed cloud only.
Try bochka
Install the CLI, then run bochka new.
Product preview