Hey HN! I built Browser AI after getting frustrated with the official Playwright MCP server in Claude Code. If you've used it, you know the pain: - 20+ separate tools that Claude has to juggle (click, fill, navigate, wait_for_selector, etc.)
- Context window explosions from verbose screenshot data and DOM snapshots
- Constant tool selection errors where Claude picks the wrong tool or forgets parameters
Browser AI wraps Playwright MCP with a single natural language interface. Instead of Claude managing dozens of tools, you get one: browser_ai.execute. Just tell it what you want in plain English: "Go to news.ycombinator.com and click on the first comment link"
"Fill out the login form with test@example.com and submit"
"Take a screenshot of the pricing table"
The magic: Browser AI uses Claude (via your Claude Code subscription) to interpret your commands and translate them into the right Playwright calls. This means: - 80% less context usage: No more massive tool definitions or snapshot data in your conversation
- Actually works: The subagent handles tool selection, so Claude doesn't have to
- OAuth auth: Uses your existing Claude Code subscription, no separate API keys needed
- Smart error recovery: If something fails, you get helpful feedback, not cryptic Playwright errors
Technical details: - Built with TypeScript and the MCP SDK
- Uses AI SDK
- Open source (MIT)
Installation is two lines:npx @qckfx/browser-ai --auth claude mcp add browser-ai -- npx --yes @qckfx/browser-ai@latest I've been using this for web scraping and E2E test writing, and it's completely changed how I work with browser automation in Claude Code. Would love to hear if this solves similar pain points for others! GitHub: https://github.com/qckfx/browser-ai npm: https://www.npmjs.com/package/@qckfx/browser-ai |