browser-use
Verified · 33 days agoControl a real Chrome browser to complete any task: fill forms, extract data, book flights.
Structured web intelligence for AI agents. 5x fewer tokens than Playwright/Firecrawl, 84.7% accuracy
claude mcp add pagemap -- uvx retio-pagemap
Python MCP server (plus SDK and CLI) that drives Chromium via Playwright and compresses pages into 2–5K-token structured page maps, with 13 tools covering clicks, form-filling, tabs, scrolling, and screenshots; 16 auto-detected page types, cookie-consent auto-dismiss across 7 CMPs, bot-detection reporting, and extraction rules for 30+ e-commerce sites. The headline numbers — 84.7% task success and 5x fewer tokens than Playwright MCP — come from the vendor's own 94-task benchmark, so treat them as self-reported. It is AGPL-3.0 (copyleft, despite the registry's NOASSERTION), adoption is thin (~75 PyPI downloads/week, zero watchers), and commits stopped in late May 2026. Interesting for token-conscious shopping and e-commerce agents; verify the claims on your own workloads.
Control a real Chrome browser to complete any task: fill forms, extract data, book flights.
Web scraping with stealth HTTP, real browsers, and Cloudflare bypass. CSS selectors supported.
MCP server for browser use access
Playwright Tools for MCP
AI browser agent: browses, clicks, types, and reads real web pages from plain-English instructions.
AI browser agent: browses, clicks, types, and reads real web pages from plain-English instructions.