Repolode
Submit
steel-dev

steel-dev/steel-browser

Open-source browser API for AI agents and apps, running a managed Chrome sandbox you can drive with Puppeteer, Playwright, or Selenium.

Open-source alternative to:Browserbase

View on GitHubWebsite
Stars
7,845
Forks
997
Open issues
76
Last commit
11 hours ago

What is it?

Steel is a browser API that lets AI agents and applications interact with live web pages without building their own browser automation infrastructure. It handles browser sessions, pages, and browser processes, so developers can focus on the application logic while the browser runs in the background.

Clients connect over Chrome DevTools Protocol using Puppeteer, Playwright, or Selenium. The API supports session options such as proxy configuration, custom Chrome extensions, viewport dimensions, and ad blocking, and sessions keep cookies and local storage between requests. A built-in UI lets you view and debug sessions, and request logging is included.

Beyond sessions, Steel exposes quick endpoints for scraping pages to clean content, taking screenshots, and generating PDFs. It can run as a Docker container locally or be self-hosted, and a hosted cloud service is also offered.

Key features

  • 01Control Chrome via Puppeteer, Playwright, or Selenium over CDP
  • 02Persistent sessions that keep cookies and local storage across requests
  • 03Proxy configuration and custom Chrome extensions per session
  • 04Built-in UI and request logging for debugging sessions
  • 05Endpoints to scrape pages, take screenshots, and generate PDFs

Use cases

  • Giving an AI agent a live browser to navigate and act on websites
  • Running browser automation jobs without managing Chrome yourself
  • Scraping pages into clean content for LLM pipelines
  • Generating screenshots or PDFs of web pages on demand

Quick start

  1. 1

    Run the pre-built Docker image locally

    docker run -p 3000:3000 -p 9223:9223 ghcr.io/steel-dev/steel-browser
  2. 2

    Open the UI at localhost:3000/ui and create a session using the API on port 3000

Things to know

  • The README describes Steel as in public beta and changing often.
  • Running locally requires Docker, or Node.js with Chrome installed.

Topics

#browser-automation#ai-agents#puppeteer#playwright#web-scraping#headless-browser#selenium#ai#ai-tools#llm

FAQ

What is steel-browser?
Open-source browser API for AI agents and apps, running a managed Chrome sandbox you can drive with Puppeteer, Playwright, or Selenium. Steel is a browser API that lets AI agents and applications interact with live web pages without building their own browser automation infrastructure. It handles browser sessions, pages, and browser processes, so developers can focus on the application logic while the browser runs in the background.
Is steel-browser free to use? What license does it have?
steel-browser is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is steel-browser still maintained?
Actively maintained. The last commit to steel-browser was on Oct 20, 2026.
What can I use steel-browser for?
Giving an AI agent a live browser to navigate and act on websites Running browser automation jobs without managing Chrome yourself Scraping pages into clean content for LLM pipelines Generating screenshots or PDFs of web pages on demand
How do I install steel-browser?
Run the pre-built Docker image locally: docker run -p 3000:3000 -p 9223:9223 ghcr.io/steel-dev/steel-browser Open the UI at localhost:3000/ui and create a session using the API on port 3000
Where can I get steel-browser?
The source code and installation instructions are on GitHub: https://github.com/steel-dev/steel-browser

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects