Browser Plus

Intelligent browser automation that routes between OpenClaw native browser tool and Vercel agent-browser based on element type detection.

Overview

Browser Plus provides a unified API for browser automation that intelligently selects the best execution strategy:

• Rich text editors (contenteditable, tweet textarea, etc.) → Uses agent-browser keyboard simulation for real key events
• Plain inputs (text fields, password fields, etc.) → Uses native browser fill for faster execution

Quick Start

const browserPlus = require('./skills/browser-plus');

// Auto-routes based on element type
await browserPlus.type({ ref: '@e12', text: 'Hello World' });

// One-click Twitter posting
await browserPlus.tweet({ text: 'My automated tweet!' });

Core API

type(options)

Types text into an element with automatic routing.

await browserPlus.type({
  ref: '@e12',           // Element reference (required)
  text: 'Hello World',   // Text to type (required)
  forceNative: false,    // Force native browser (optional)
  forceAgent: false      // Force agent-browser (optional)
});

click(options)

Clicks an element.

await browserPlus.click({
  ref: '@e12',           // Element reference (required)
  forceAgent: false      // Use agent-browser (optional)
});

Composite Actions

tweet(options)

One-click Twitter posting.

await browserPlus.tweet({
  text: 'My tweet content',
  media: ['/path/to/image.jpg']  // Optional media attachments
});

How It Works

1. Element Detection: Analyzes the target element to determine its type
2. Strategy Selection:
   • Rich text editors → agent-browser (real key events)
   • Plain inputs → native browser (faster)
3. Execution: Routes to the appropriate adapter

Scripts

• scripts/detectors.js - Element type detection utilities
• scripts/adapters/native-browser.js - OpenClaw browser wrapper
• scripts/adapters/agent-browser.js - Vercel agent-browser CLI wrapper
• scripts/composite/tweet.js - Twitter posting composite action