PinchTab with OpenClaw

PinchTab is a standalone HTTP server that gives AI agents direct control over Chrome. It delivers roughly 800 tokens per page with text extraction—5–13x cheaper than screenshots—and supports multi-instance (parallel Chrome processes with isolated profiles). The binary is a 12MB Go executable, suitable for headless or headed mode and Raspberry Pi (ARM64).

Run PinchTab as a separate service; the OpenClaw pinchtab skill lets your agent navigate, snapshot, click, fill, and extract via HTTP. Best for token-conscious automation when the built-in browser is too heavy or expensive.

curl -fsSL https://pinchtab.com/install.sh | bash
# or
npm install -g pinchtab

Docs: pinchtab.com/docs. GitHub: pinchtab/pinchtab.

• Token-efficient — ~800 tokens/page vs thousands for screenshots
• Multi-instance — Isolated Chrome profiles for parallel workflows
• Small binary — 12MB Go, no Node runtime
• ARM64 support — Runs on Raspberry Pi
• HTTP API — Integrates with OpenClaw via pinchtab skill

Use PinchTab when you need token-conscious, repeated page visits—forms, monitoring, data extraction. Use the built-in browser when you need full DOM access, Canvas integration, or one-off exploratory browsing. For protected sites (403, Cloudflare), see GhostFetch and camofox-browser.