OpenHuman Guide
← Back to Guides

OpenHuman Web Scraper & Browser Control — Browsing and Desktop Automation

OpenHuman ships with two powerful web tools: Web Scraper extracts text content from any URL; Browser & Computer Control operates your browser and desktop like a human. Together, they handle everything from information extraction to interactive workflows.

🌐 Web Scraper

When you only need the page content without interaction, Web Scraper is the direct approach.

  • Extracts plain text from any URL
  • Auto-converts HTML to structured Markdown
  • Output compressed via TokenJuice
  • Results can be stored in Memory Tree

🖥️ Browser & Computer Control

When the AI needs to operate your machine like a person — open pages, screenshot, click, type.

Browser

  • Open a URL in embedded webview
  • Screenshot the current page
  • Inspect image output and metadata

Mouse + Keyboard

  • Mouse: move, click, drag
  • Keyboard: type text, send key chords
  • Human-like trajectories avoid bot detection

📊 When to use which

ScenarioToolWhy
Read article/docsScraperFaster, fewer tokens
Fill forms / loginBrowser controlNeeds interaction
Monitor page changesScraperPeriodic fetch
Multi-step UI flowBrowser controlNeeds chained actions

🔒 Safety

  • Browser runs in CEF sandbox — no access to personal browser data
  • Filesystem tools scoped to workspace
  • All tool output compressed via TokenJuice