IMPORTANT: THE MASTER AGENT IS A LOCAL AGENT! BE PRECISE WHEN ASKING QUESTIONS. Local small models have limitations with function calling and MCP protocol interpretation - they require clear, direct prompts to properly route requests to specialized agents.
A multi-agent system implementing the Multi-agent Conversation Protocol (MCP) with various specialized server components.

- ML/DL News Server - Provides tools for machine learning and deep learning news search
- Gmail Server - Enables email interaction via Gmail
- Voice Call Server - Facilitates voice calls and transcriptions
- CAG+RAG Server - Provides context-aware generation and retrieval augmented generation
- Video Generator - Creates videos with customizable content and music
- Create a virtual environment:
uv venv - Activate the environment:
source .venv/bin/activate - Install dependencies:
uv sync - Configure required environment variables:
VOICE_SERVER_URL- Required for voice call functionalityOPENAI_API_KEY- Required for CAG+RAG and video generationREPLICATE_API_KEY- Required for video generationSONAUTO_API_KEY- Required for video audio generation
Two interfaces are available:
- Terminal Interface:
uv run agent.py- Command-line interaction with all agents - Web UI Interface:
uv run web_ui.py- Stylish web-based chat interface at http://localhost:5000
Note: Make sure you have Ollama running if using the default agent configuration.
- Modern, responsive chat interface
- Real-time status indicators
- Example query suggestions
- Support for code blocks and formatting
- Works with all MCP agent capabilities
The Video Generator agent creates custom AI-generated videos with specified content and music. It uses:
- Replicate for video frame generation
- SonAuto for background music and audio synthesis
- OpenAI for content creation and scripting
Example video generated with the prompt "Create a video about a futuristic programming school with rock music":
Click the image above to watch the video on YouTube.
- Ask about ML/DL news: "What are the latest developments in transformer models?"
- Use Gmail: "Check my recent emails" or "Send an email to example@example.com"
- Make voice calls: "Call +1234567890 and ask about meeting availability"
- Create videos: "Create a video about a futuristic programming school with rock music"
- Query the codebase: "How does the voice call system work?"




