ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified browser automation Safety 3/5

pinchtab

通过 PinchTab HTTP API 控制无头或有头 Chrome 浏览器,用于网页自动化、爬虫、表单填充、导航、截图和数据提取

Why use this skill?

Control Chrome with AI using PinchTab. A lightweight tool for automated web navigation, form filling, and data extraction via a clean accessibility tree API.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/hellotombruce/pinchtab-skills
Or

What This Skill Does

PinchTab is a high-performance browser automation tool for AI agents, enabling direct interaction with Chrome via an HTTP API based on the accessibility tree. Unlike traditional Selenium or Playwright setups that can be heavy, PinchTab provides a lightweight, token-efficient interface that allows AI agents to 'see' and interact with the web just like a human. It abstracts the complexity of DOM trees into stable, concise references (refs), making it ideal for autonomous navigation, form submission, data scraping, and complex multi-step web workflows.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/hellotombruce/pinchtab-skills

Ensure you have Chrome installed on your machine. Once installed, you can start the PinchTab daemon using the pinchtab command, and configure it using environment variables like BRIDGE_PORT or BRIDGE_TOKEN for remote access control.

Use Cases

  • Automated Research: Direct your agent to search Google, click through search results, and extract relevant text to build summary reports.
  • Dynamic Form Filling: Use agents to register for services, sign up for newsletters, or fill out complex multi-page surveys by programmatically interacting with form inputs.
  • Data Extraction: Scrape structured data from difficult-to-parse websites by navigating through pages and capturing the accessibility tree in compact formats.
  • Automated Testing: Trigger regression tests on web applications by simulating user clicks, key presses, and navigation patterns.

Example Prompts

  1. "PinchTab, please navigate to the OpenClaw documentation site, find the search bar, type 'installation' and press Enter. Then extract the titles of the first three search results."
  2. "Go to the following URL [link] and fill out the contact form with these details: name 'AI Agent', email '[email protected]', and submit it."
  3. "Open a new tab, navigate to the news website, take a full-page screenshot and save it as landing_page.jpg for my review."

Tips & Limitations

  • Security First: Always use a clean, dedicated Chrome profile. Never link your personal browsing profile to PinchTab to prevent unauthorized access to your saved cookies and credentials.
  • Token Optimization: Always use the compact format when requesting snapshots to save context tokens, as this removes redundant metadata and focuses on interactive elements.
  • Stability: Use the provided reference IDs (e.g., e5, e12) for actions rather than relying on dynamic DOM paths, as these references are designed to be persistent across most page states.
  • Limitations: The agent can only see what is visible in the current viewport or accessibility tree. Complex shadow DOMs or canvas-based elements may require custom evaluation scripts.

Metadata

Stars2387
Views0
Updated2026-03-09
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-hellotombruce-pinchtab-skills": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags(AI)

#browser-automation#web-scraping#chrome-control#agent-tools
Safety Score: 3/5

Flags: network-access, file-write, file-read, code-execution