Skip to content

Latest commit

 

History

History
77 lines (54 loc) · 2.28 KB

README.md

File metadata and controls

77 lines (54 loc) · 2.28 KB

ProAction - Proactive AI Assistant for AugmentOS

ProAction is a multi-agent, proactive AI assistant that monitors and analyzes speech transcriptions to provide relevant insights, factual information, and assistance in real-time through AugmentOS smartglasses.

Features

  • Multi-agent Architecture: Extensible system of specialized AI agents that each focus on different types of assistance
  • Priority System: Configurable priority levels to control when and how information is presented
  • Contextual Awareness: Maintains short-term and long-term memory to provide relevant information
  • Voice Commands: Natural language interface for controlling the system
  • Configurable: User-configurable agent settings and priority thresholds

Agents

Currently Implemented:

  • Factual Agent: Provides factual information and verifies statements

Planned:

  • Quiz Agent: Helps with studying and recalling information
  • Personal Assistant: Manages reminders, tasks, and calendar events

Usage

Installation

  1. Set up environment variables:

    AUGMENTOS_API_KEY=your_api_key
OPENAI_API_KEY=your_openai_key
PORT=3000 (optional)
AUGMENTOS_WS_URL=wss://staging.augmentos.org/tpa-ws (optional)

  2. Install dependencies:

    npm install

  3. Build and start the server:

    npm run build
npm start

Voice Commands

  • "Show insights" - Display pending insights
  • "Hide insights" - Clear the current display
  • "Show agent settings" - View and configure agents
  • "Enable [agent name]" - Enable a specific agent
  • "Disable [agent name]" - Disable a specific agent
  • "Help me" - Show help information

Architecture

ProAction uses a coordinator-based architecture where:

  1. Transcriptions are received from AugmentOS
  2. The coordinator processes the transcription
  3. Active agents analyze the transcription and generate insights
  4. Insights are prioritized and presented to the user based on configured thresholds

Development

Adding New Agents

  1. Create a new agent class in src/agents/ that extends BaseAgent
  2. Implement the required methods
  3. Register the agent in src/server.ts

Customizing Priority Levels

Modify the thresholds in the ConfigManager to adjust when different insights are shown.

License

MIT