Best forRunning many headless browsers in the cloud with session recordings, proxies and captcha handling.
Trade-offHosted only, billed by browser time.
Used by 16 makers' products for this
For browser automation for agents, the alternatives to Steel that makers' products we track chose most are Browserbase (16), Browser Use (12) and Kernel (1). 6 of the 13 are open source or self-hostable (Browser Use, Stagehand, Playwright and 3 more) and 3 have a free tier.
Not a ranking: each option is described by the situation it fits, with its trade-off.
How many makers' products we track use each for browser automation for agents.
| Option | Kind | Makers' products | Free tier | Open source | |
|---|---|---|---|---|---|
| Hosted browsers | 0 | — | Yes | ||
| Hosted browsers | 16 | — | No | ||
| Agent library | 12 | — | Yes | ||
| Agent library | 0 | — | Yes | ||
| Automation library | 0 | — | Yes | ||
| Automation library | 0 | — | Yes | ||
| Hosted browsers | 1 | — | No | ||
| Agent library | 1 | Yes | Yes | ||
| Hosted browsers | 1 | Yes | Self-hostable | ||
| Hyperbrowser | Hosted browsers | 0 | — | No | |
| Agent library | 0 | Yes | No | ||
| Agent library | 0 | No | No | ||
| Hosted browsers | 0 | — | No | ||
| Agent library | 0 | — | No |
Browser infrastructure or agent logic. Browser infrastructure runs the headless browsers for you, with sessions, proxies and recordings, and your code decides what to do. Agent libraries supply the model-driven logic that decides where to click, and can run on your own browser or a hosted one. Underneath both sits a plain automation library, where you script every step yourself.
Steel (Hosted browsers): Browser infrastructure for agents that you can use as a cloud service or self-host from open source. Trade-off: Smaller ecosystem than larger browser-infrastructure providers.
Best forRunning many headless browsers in the cloud with session recordings, proxies and captcha handling.
Trade-offHosted only, billed by browser time.
Used by 16 makers' products for this
Best forGiving a model a browser and a goal in a few lines of Python, and letting it work out the clicks.
Trade-offEvery step costs model calls and time, and results vary run to run.
Used by 12 makers' products for this
Best forMixing scripted browser steps with natural-language act, extract and agent calls, in TypeScript or Python.
Trade-offWorks best on Browserbase, and each natural-language step still costs a model call.
No maker product tracked for this yet
Best forScripting Chromium, Firefox and WebKit with one API, in TypeScript, Python, Java or .NET.
Trade-offScripts break when page structure changes, and you run the browsers yourself.
No maker product tracked for this yet
Best forControlling headless Chrome from Node.js, for scraping, screenshots and PDF generation.
Trade-offJavaScript only, with fewer built-in waits and test tools than Playwright.
No maker product tracked for this yet
Best forCloud browsers for agents with reusable sessions and handling for bot checks; the browser images are open source.
Trade-offThe browser images are open source, the hosted control plane is not.
Used by 1 maker's product for this
Best forAutomating form filling, logins and downloads on websites the agent hasn't seen before, open source or as a hosted cloud with CAPTCHA solving and 2FA.
Trade-offAGPL-3.0 licence if you self-host and modify it; the cloud bills in credits per run.
Used by 1 maker's product for this
Best forPointing existing Puppeteer or Playwright code at hosted browsers, plus REST endpoints for screenshots, PDFs and scraping.
Trade-offSelf-hosting a closed-source product needs a commercial licence (the code is SSPL); usage is billed in units.
Used by 1 maker's product for this
Best forCloud Chrome sessions plus scraping and crawling APIs from one vendor.
Trade-offCredit-based billing across products; free 5,000 credits, then $30 or $100 a month.
No maker product tracked for this yet
Best forFinding page elements and pulling structured data with natural-language-like queries in place of CSS or XPath selectors, on Playwright or its remote browsers.
Trade-offEach query is a paid API call after the small free allowance; your code still drives the steps.
No maker product tracked for this yet
Best forLetting Claude itself drive a browser or desktop through the API's browser-use and computer-use toolsets, while your application runs the browser and carries out each action.
Trade-offYou run the browser or desktop environment yourself; Anthropic notes latency is slow next to a human and recommends a sandboxed VM or container.
No maker product tracked for this yet
Best forLetting ChatGPT subscribers hand web tasks to an agent that drove its own browser.
Trade-offFolded into agent mode in ChatGPT; it never had an API.
No maker product tracked for this yet
Best forCalling an API that ran a web-browsing agent to complete a task for you.
Trade-offPivoted to AGI, Inc. (on-device action AI for Android); its SDK has had no release since 2024-10-08.
No maker product tracked for this yet
It depends on what you need it for; we don't rank them. The one makers' products we track chose most for browser automation for agents is Browserbase (16 makers' products). The agent must handle pages it has never seen, given only a goal: Browser Use. You know most of the steps and want the model only for the parts that vary: Stagehand. The flow is fixed and can be scripted without a model at all: Playwright and Puppeteer.
Yes. For browser automation for agents, Skyvern, Browserless and AgentQL have a free tier a small product can start on.
Yes. For browser automation for agents: Browser Use, Stagehand, Playwright, Puppeteer, Skyvern and Browserless (self-hostable).