Steel

• ▸Sessions API: create and control isolated cloud browser sessions programmatically
• ▸Anti-bot tooling: stealth settings, residential proxies, CAPTCHA solving, mobile-mode browsing
• ▸Persistent Profiles API: reuse auth, cookies, extensions, and credentials across sessions
• ▸Content extraction: return page content as HTML, markdown, cleaned HTML, or readability text via a scrape endpoint
• ▸Files, Extensions, and Credentials APIs for managing session state and Chrome extensions
• ▸Observability: live sessions, past-session embeds, recordings, WebRTC streaming, and agent logs
• ▸Framework/tool integrations: Playwright, Puppeteer, Selenium, Browser-Use, Magnitude, CrewAI, and Claude/Gemini/OpenAI computer-use
• ▸Native Node, Python, Rust, and Go SDKs, plus a CLI and an installable coding-agent skill (steel-browser)

Authentication

API key sent via the steel-api-key header, or STEEL_API_KEY environment variable for the SDKs

Base URL

https://api.steel.dev

Notable endpoints

• POST /v1/scrape - scrape a URL without a browser session, returning HTML/markdown/cleaned HTML/readability text
• sessions.create() - create a cloud browser session (Node/Python SDK)
• sessions.release(sessionId) - release an active session
• for Playwright/Puppeteer over CDP
• Captchas API - detect and solve CAPTCHAs during automation
• Profiles API - reuse auth, cookies, extensions, and credentials across sessions
• Files API - upload, download, and manage files during a session
• Extensions API - add Chrome extensions to a session
• Credentials API - store credentials for users or agents

Machine-readable spec

OpenAPI spec retrieved

Published by the vendor themselves — not community forks.

• steel-browser

  Installable coding-agent skill that gives an agent control of a Steel browser session, installed via the Steel CLI.

It exposes cloud sessions through a REST API with a published OpenAPI spec, is self-hostable via Docker or a one-click Railway deploy, and ships a CLI plus a dedicated agent skill package for coding agents.

AI agent builders needing reliable, authenticated browser access in production; scraping/data-extraction teams that don't want to run headless Chrome infrastructure themselves; and teams migrating existing Playwright/Puppeteer/Selenium scripts to managed cloud browsers.

Teams wanting a fully managed, no-code scraping product rather than developer infrastructure are a mismatch: Steel is explicitly an API/CLI/SDK surface for builders, not an end-user scraping app.

No curated rival set covers this tool yet, so these are simply other tools in the same architectural layer and product form. Treat it as a starting point, not a shortlist.

• Pricing plans, billing model, and free-tier limits
• Whether Steel exposes an MCP server (no /mcp page found)
• LinkedIn, X, YouTube social links
• Published pricing was not available in the reviewed source set; confirm current commercial terms with the vendor.