OpenHuman Native Tools — Browser Control, Web Scraper, Coder Toolset, Cron

OpenHuman ships with a complete set of native tools, ready to use out of the box. No plugins or extensions required.

🔍 Web Search

Built-in web search (via managed backend proxy). The AI can fetch real-time information beyond its training data cutoff.

• Auto-invokes search engines for results
• Output compressed via TokenJuice before entering model context
• Managed backend by default; can configure a custom search API

🌐 Web Scraper

Built-in web fetching tool that converts HTML to structured Markdown for AI analysis.

• HTTP GET request to any URL
• Auto-converts HTML to structured Markdown
• Output passes through TokenJuice compression
• AI reads and understands web content naturally

🖥️ Browser & Computer Control

OpenHuman can control your browser and desktop for complex interactive tasks.

• Open pages, click, fill forms, scroll, screenshot
• Read DOM structure and content
• Execute JavaScript
• Simulate human-like mouse movements and keystrokes

Example: "Star this repo on GitHub" → AI opens browser, navigates, clicks the star button.

💻 Coder Toolset

⏰ Cron & Scheduling

Built-in cron system. Set scheduled tasks directly from conversation:

• cron-like syntax for recurring schedules
• One-shot timer reminders
• Periodic execution (daily, weekly, monthly)

Example: "Give me an email digest every morning at 9 AM" → AI creates a cron job, executes automatically.

🔁 Memory Tools

AI can read, write, search and manage Memory Tree data:

• Search historical memories
• Write new memory entries
• Organize and link related memories
• Delete or correct inaccurate memories

👥 Agent Coordination

OpenHuman supports spawning sub-agents for parallel task execution:

• Main agent delegates subtasks to sub-agents
• Sub-agents work independently and report back
• Ideal for complex multi-step workflows

Example: "Analyze this 100-page document — you and another agent each read 50 pages" → parallel execution.

📊 Toolchain Comparison