Integrates with Google's Generative AI (Gemini) through API keys for natural language processing capabilities
Provides access to Google Maps functionality through a dedicated toolset, allowing the agent to perform location-based queries and mapping operations
ADK MCP Agent
A streaming chat agent using Google ADK and the Model Context Protocol (MCP) Google Maps toolset.
Local Development
Install dependencies:
pip install -r requirements.txtSet up
.envinapp/with your API keys.Run locally:
export SSL_CERT_FILE=$(python -m certifi) uvicorn app.main:app --reload
Related MCP server: ChatterBox MCP Server
Project Structure
The agent logic is in the
agent/directory (agent/agents.py).The FastAPI app and UI are in the
app/directory.
Deployment
To test and run the app locally with the UI:
Install dependencies:
pip install -r requirements.txtSet up environment variables:
Create a
.envfile inside theapp/directory with your API keys:GOOGLE_GENAI_USE_VERTEXAI=FALSE GOOGLE_API_KEY=your_genai_api_key GOOGLE_MAPS_API_KEY=your_google_maps_api_key
Run the FastAPI app:
export SSL_CERT_FILE=$(python -m certifi) uvicorn app.main:app --reloadAccess the UI:
Open your browser and go to http://localhost:8000
You can now interact with the chat interface and test the agent locally.
License
MIT