Contact Book MCP Server

A simple Model Context Protocol (MCP) server that manages an in-memory contact book. Built with FastMCP from the mcp Python SDK.

Features

This server exposes four tools:

Tool	Description
add_contact(name, phone, email)	Adds a new contact to the contact book.
get_contact(name)	Retrieves a contact by exact name match.
search_contacts(query)	Searches contacts by partial (case-insensitive) name match.
update_contact(name, phone, email)	Updates an existing contact's phone and email.

Contacts are stored in memory as:

contacts: dict[str, dict]  # name -> {"phone": ..., "email": ...}

Note: Data is stored in memory only. All contacts are lost when the server restarts.

Requirements

• Python 3.10+
• uv (recommended for running the server)

Installation

1. Clone this repository:

   git clone <your-repo-url>
   cd contact_book_server_mcp
   

2. Install dependencies:

   pip install -r requirements.txt
   

   Or, if using uv (recommended, no manual install needed — see below).

Running the Server

With MCP Inspector (for testing/development)

uv run --with mcp mcp dev server.py

This starts the server and opens the MCP Inspector in your browser, where you can test each tool interactively.

Standalone

python server.py

Example Usage

1. Add a contact

   add_contact(name="Humna Fahad", phone="03001234567", email="humna@test.com")
   → "Contact 'Humna Fahad' added successfully."
   

2. Get a contact

   get_contact(name="Humna Fahad")
   → {"name": "Humna Fahad", "phone": "03001234567", "email": "humna@test.com"}
   

3. Search contacts

   search_contacts(query="Hum")
   → [{"name": "Humna Fahad", "phone": "03001234567", "email": "humna@test.com"}]
   

4. Update a contact

   update_contact(name="Humna Fahad", phone="03009998888", email="humna@updated.com")
   → "Contact 'Humna Fahad' updated successfully."
   

Project Status

✅ All four tools implemented and tested via MCP Inspector.

Possible Improvements

• Add a delete_contact tool.
• Add a list_all_contacts tool.
• Persist contacts to a JSON file or database instead of in-memory storage.