adk-mcp-servers

Hamza-Chergui-Converteo/adk-mcp-servers

3.2

If you are the rightful owner of adk-mcp-servers and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

This project demonstrates the implementation of a Model Context Protocol (MCP) server using a SQLite CRM database, integrated with a Google ADK-powered agent.

Tools
6
Resources
0
Prompts
0

MCP CRM Example with Google ADK

This repository demonstrates how to build an MCP (Model Context Protocol) server backed by a SQLite CRM database, and how to interact with it using a Gemini-powered agent via the Google Application Development Kit (ADK).

🚀 What is This Project?

  • ✅ A ready-to-use MCP server using a SQLite CRM database
  • 🤖 A custom Google ADK agent capable of querying, updating, and analyzing CRM data
  • 💬 A simulated CLI chat interface to test interactions with the agent
  • 🧰 A toolset of 6 MCP tools exposed to the agent (read, write, create, list tables, describe schema, add insights)

🧱 Project Structure

mcp-crm-project/
├── agent.py           # Main agent logic (Gemini + ADK)
├── db/
│   ├── setup_db.py    # Script to initialize the SQLite database
│   └── crm.db         # Auto-generated SQLite CRM database
├── sqlite/            # MCP server adapted from official repo
└── README.md          # You're here!

🛠️ Setup Instructions

1. Clone the Repository

git clone https://github.com/Hamza-Chergui-Converteo/adk-mcp-servers.git
cd mcp-crm-project

2. Initialize the SQLite Database

cd db
python setup_db.py

This will create a crm.db file with mock data.

3. Add Your Gemini API Key

In agent.py, replace YOUR_API_KEY with your actual Google API key:

os.environ["GOOGLE_API_KEY"] = "your-key-here"

🧠 Agent Toolset

The MCP server exposes 6 tools accessible by the agent:

🔍 Query Tools

  • read_query: Run SELECT queries
  • write_query: INSERT, UPDATE, DELETE
  • create_table: Create new tables

🧾 Schema Tools

  • list_tables: List all tables
  • describe_table: Get schema of a table

📊 Insight Tools

  • append_insight: Add business insights to memo://insights

💬 Start Chatting with the Agent

Simply run:

python agent.py

You’ll enter a CLI chat with the CRM agent. Type your questions or SQL commands (e.g., “list all prospects”, “add a new contact”, etc.). Type exit to quit.

📚 Reference

This project is based on the official MCP servers repository:
🔗 https://github.com/modelcontextprotocol/servers

Feel free to explore it for other backends (Redis, OpenAI, Slack, etc.).