crm-mcp-ai-agent
This project aims to integrate an AI agent into a CRM system. It is designed to analyze customer data and provide automated responses, thereby improving operational efficiency. Implemented in Python, it offers an easy-to-use API for integration with other systems.
GitHub Stars
0
User Rating
Not Rated
Favorites
0
Views
32
Forks
0
Issues
0
Agentic AI with Model Context Protocol (MCP)
An interactive AI assistant powered by PydanticAI and the Model Context Protocol (MCP). The agent can perform various tasks using available tools through a simple chat interface.
Screenshots
Click to view CRM - Add Customer
Click to view CRM - List Customers
Click to view Agent Tool Usage
*Agent using tool, it also shows which tools did it use and arguments. Then you can see the response from MCP
Click to view Weather Tool
Click to view Tool Fetching Process
Click to view Tool Fetching Extended
Features
- Time & Date: Get current time and date information
- Weather: Check weather conditions and forecasts (weatherapi.com)
- Web Search: Search the web for information (duckduckgo)
- Memory: Store and recall information across conversations
- Planning: Create task plans and manage goals
- Calculator: Perform mathematical calculations
- CRM: Manage customer information (MongoDB)
Setup
Install dependencies:
conda env create -f environment.yml conda activate mcp-agentCreate
.envfile with your API configuration:API_KEY=your_api_key_here MCP_SERVER_URL=http://localhost:8001/sse BASE_URL=https://api.mistral.ai/v1MONGO_URI=mongodb+srv://<password>@<atlas-url>.mongodb.net/?retryWrites=true&w=majority&appName=Cluster0 MONGO_DB=June16db MONGO_COLLECTION=customers WEATHER_API_KEY=<https://www.weatherapi.com/>
Usage
Start the MCP server:
fastmcp run server/main.py --transport sse --port 8001Start the web interface (in a new terminal):
chainlit run app.pyOpen your browser and go to
http://localhost:8000
Example Queries
- "What time is it?"
- "What's the weather in Paris?"
- "Remember that I have a meeting at 3 PM"
- "Calculate 15 * 24 + 100"
- "Create a plan for learning Python"
- "Search for information about AI"
Manual Architecture (no longer using, use Pydantic instead)
┌────────────────┐ ┌─────────────────┐ ┌──────────────┐
│ │ │ │ │ │
│ Chainlit UI │◄───►│ Handlers.py │◄───►│ LLM API │
│ (app.py) │ │ │ │ │
│ │ │ │ │ │
└────────────────┘ └────────┬────────┘ └──────────────┘
│
▼
┌─────────────────┐
│ │
│ MCP Client │
│ (mcp_client.py)│
│ │
└────────┬────────┘
│
▼
┌─────────────────┐
│ │
│ MCP Server │
│ (mcp_server.py)│
│ │
└─────────────────┘
Pydantic AI is like combination of MCP Client and Handler
License
MIT License
0
Followers
0
Repositories
0
Gists
0
Total Contributions
MCP server built with Google’s Agent Development(ADK) Kit featuring multiple specialised agents that collaborate through intelligent routing. The system provides RESTful API endpoints for various capabilities including web searches, data analysis, and code generation while maintaining extensibility through a modular architecture