Exposes FastAPI application endpoints as MCP tools, enabling AI agents to interact with REST API functionality including user management, task management, dice rolling, and application health monitoring
Integrates with Google's Gemini AI model to enable natural language interaction with MCP server tools, allowing conversational access to FastAPI endpoints and application functionality
FastAPI + MCP + Gemini Integration
This project demonstrates how to integrate a FastAPI application with Google's Gemini AI using a simplified MCP (Model Context Protocol) server implementation.
๐๏ธ Architecture
FastAPI App (
app.py
): A sample REST API with user management, task management, and dice rollingSimple MCP Server (
simple_mcp_server.py
): Simplified MCP server that exposes FastAPI endpoints as toolsGemini Integration (
simple_gemini_integration.py
): Connects Gemini AI with the MCP server
๐ Features
FastAPI Application
User management (CRUD operations)
Task management with completion tracking
Dice rolling functionality
Health checks and statistics
RESTful API endpoints
MCP Server Tools
get_health_status()
: Check application healthget_app_info()
: Get application informationget_all_users()
: Retrieve all userscreate_user()
: Create new usersget_user_by_id()
: Get specific userget_all_tasks()
: Retrieve all taskscreate_task()
: Create new taskscomplete_task()
: Mark tasks as completedroll_dice()
: Roll dice with custom parametersget_app_statistics()
: Get application statisticssearch_users_by_name()
: Search users by nameget_pending_tasks()
: Get incomplete tasksget_completed_tasks()
: Get completed tasks
๐ Prerequisites
Python 3.8+
Google Gemini API key (optional - demo works in simulation mode)
Basic Python packages (fastapi, uvicorn, aiohttp, google-generativeai)
๐ ๏ธ Installation
Clone or download the project files
Install dependencies:
pip install fastapi uvicorn aiohttp google-generativeai python-dotenv requestsSet up environment variables (optional): Create a
.env
file and add your Gemini API key:GEMINI_API_KEY=your_actual_api_key_hereNote: The demo works without an API key in simulation mode.
Get a Gemini API key:
Visit Google AI Studio
Create a new API key
Add it to your
.env
file
๐ฏ Usage
0. Look for the video demo
You can look for the zip file in which screen recording is present. That includes a demo question and an answer.
1. Start the FastAPI Server
The FastAPI server will run on http://localhost:8000
2. Test the FastAPI Endpoints
You can test the API directly:
3. Run the Gemini Integration
Demo Mode (Predefined Queries)
Interactive Mode
Automated Demo
4. Example Gemini Queries
In interactive mode, you can ask questions like:
"Check the health status of the FastAPI application"
"Create a new user named 'Alice' with email 'alice@example.com' and age 25"
"Create a task called 'Learn Python' with description 'Study Python programming'"
"Roll 5 dice with 10 sides each"
"Show me all users and get the application statistics"
"Mark the first task as completed"
"Show me all pending tasks"
๐ง Configuration
FastAPI Server
Default port: 8000
Host: 0.0.0.0 (accessible from all interfaces)
Modify
app.py
to change these settings
MCP Server
Connects to FastAPI server at
http://localhost:8000
Modify
API_BASE_URL
inmcp_server.py
if needed
Gemini Integration
Uses Gemini 2.0 Flash model
Configure API key via environment variable
Modify model settings in
gemini_integration.py
๐ Project Structure
๐งช Testing
Test FastAPI Endpoints
Test MCP Server
Test Gemini Integration
Test Everything
๐ Troubleshooting
Common Issues
"Please set GEMINI_API_KEY environment variable"
Make sure you have a
.env
file with your API keyCheck that the API key is valid
"Error connecting to MCP server"
Ensure the FastAPI server is running on port 8000
Check that all dependencies are installed
"ModuleNotFoundError"
Run
pip install -r requirements.txt
Make sure you're using Python 3.8+
Debug Mode
To see more detailed error messages, you can modify the integration script to include more logging.
๐ Next Steps
Add more FastAPI endpoints
Create additional MCP tools
Implement authentication
Add database persistence
Create a web interface
Deploy to cloud platforms
๐ Learn More
๐ค Contributing
Feel free to submit issues and enhancement requests!
๐ License
This project is open source and available under the MIT License.
This server cannot be installed
local-only server
The server can only run on the client's local machine because it depends on local resources.
Enables Gemini AI to interact with a FastAPI application through MCP tools for user management, task management, and dice rolling functionality. Provides natural language access to REST API endpoints including CRUD operations, health checks, and application statistics.