#4545th of 82 in Agent frameworks and SDKs
pinchtab
HTTP server and CLI that lets AI agents drive Chrome with compact snapshots
- Stars
- 10.5k
- License
- MIT
- Last commit
- Sep 2026
- Last release
- Aug 2026
- Language
- Go
8,529 on 10 Apr 202610,533 on 30 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
PinchTab is a single Go binary (about 30 MB) that runs a control-plane HTTP server on port 9867 and manages headless or headed Chrome instances with persistent profiles. Agents navigate, snapshot the accessibility tree, click by stable element refs and extract text through a CLI or REST API. It also ships site audit and visual compare commands, and an agent skill for Claude, Cursor, Windsurf and Codex.
Who it is for: Developers giving local agents a browser they control through CLI or HTTP
Strengths
- Text extraction at about 800 tokens per page, versus roughly 10,000 for raw output
- Persistent named profiles keep logins across restarts; multiple isolated instances run in parallel
- Binds to 127.0.0.1 by default, with sensitive endpoints off and a local-only site allowlist
- CLI and HTTP API, plus audit and compare commands usable as a CI gate
Weaknesses
- Needs a locally installed Chrome or Chromium; none is bundled in the binary
- Not designed for multi-tenant or public-internet exposure; remote setups must be secured manually
- Default allowlist blocks public websites until the operator widens it
- Windows support is limited and less tested; the daemon workflow is not recommended there
What it needs
- no GPU
- Docker + Compose
- Needs Chrome or Chromium, Docker (optional)
- port 9867
Quick answers
Is pinchtab open source?
Yes. Its code is on GitHub under the MIT license.
Can you run pinchtab with Docker?
Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.
Does pinchtab need a GPU?
No, it runs without one.
Also in Agent frameworks and SDKs
See all 82| Rank | Project | Score |
|---|---|---|
| 1 | ppt-masterAgent skill that turns documents into natively editable PowerPoint decks | 95 out of 100 |
| 2 | openai-agents-pythonPython SDK for multi-agent workflows with handoffs, guardrails and tracing | 90 out of 100 |
| 3 | langchainPython framework for building LLM agents and apps from interoperable components | 89 out of 100 |
| #4 | browser-usePython library that lets an LLM agent control a web browser | 89 out of 100 |
| #5 | tradingagentsMulti-agent LLM framework that simulates a trading firm's analysis process | 89 out of 100 |