pinchtab vs stealth-browser-mcp
Small Go HTTP/MCP server that gives AI agents direct control over Chrome — stealth CDP injection, multi-instance orchestration and a real-time dashboard. Local-first, single binary. — versus — MCP server for undetectable browser automation: real Chrome via nodriver + CDP, Cloudflare/anti-bot bypass, AI-written network hooks — agents browse where Playwright gets blocked.
Both expose stealth real-Chrome-over-CDP to agents; stealth-browser-mcp is a Python MCP server on nodriver, PinchTab is a Go daemon adding multi-instance orchestration, an HTTP API and a dashboard.
| pinchtab | stealth-browser-mcp | |
|---|---|---|
| Stars | 9.5k | 1.5k |
| Forks | 705 | 231 |
| Language | Go | Python |
| License | MIT | MIT |
| Last activity | today | 2 months ago |
| Topics | web, agents | web |
| Curated connections | 5 | 4 |
pinchtab — the curator's take
Reach for it when you want one language-agnostic browser control plane that many agents (or non-JS stacks) share over HTTP/MCP, with stealth and several Chrome instances managed as a background daemon. Skip it if you just need an in-process Python/JS browser tool for a single agent — browser-use or browser-harness-js are lighter. The dashboard/API is a privileged operator surface: local-first by design, never expose it publicly without locking down tokens, TLS and endpoint families.
stealth-browser-mcp — the curator's take
For the pages where standard automation dies at the Cloudflare wall: real Chrome instances driven through nodriver + CDP, exposed to any MCP client, with network-hook tooling Playwright MCPs don't have. When your agent legitimately needs a protected page (your own accounts, paywalled services you subscribe to), this is the tool that actually works. The obvious caution IS the caution: 'bypasses anti-bot systems' means you're overriding sites' stated wishes — check ToS and law before pointing it anywhere you don't own; expect the arms race to break it periodically, and audit what its hooks can capture (credentials pass through).