📚 SQLite Book Library with FastMCP

This project implements a simple Book Library backend using:

• 🛠️ FastMCP: to expose Python functions as AI-callable tools
• 🗄️ SQLite: to store book data
• ⚙️ Uvicorn: to serve the tools over an HTTP SSE interface

It allows Language Models (LLMs) to interact with your local database through tool calls — ideal for Retrieval-Augmented Generation (RAG), autonomous agents, or database assistants.

🚀 Features

• 📖 Add new books using SQL
• 🔍 Query books with custom SELECT queries
• 🧠 Designed to integrate with LLM agents (e.g., via LlamaIndex)
• 🧰 Easily extendable: add update_book() or delete_book() as needed

🔧 Tech Stack

⚙️ What is MCP?

MCP (Model Context Protocol) is an open standard designed to help applications connect large language models (LLMs) to external tools and data.
Think of MCP like a USB-C port for LLMs: it standardizes the way models call functions and access data sources.

Why use MCP?

• 🔌 Tool interoperability – Easily expose Python functions as callable APIs
• 🧠 Plug-and-play LLM agents – LLMs can auto-discover and use tools
• 🔐 Vendor-agnostic – Use with OpenAI, Anthropic, Hugging Face, or local models
• 🏗️ Workflow-friendly – Ideal for building modular AI agents and pipelines

🛠️ Local MCP Setup

server.py – Your tool server

from mcp.server.fastmcp import FastMCP
mcp = FastMCP('sqlite-book-library')  # Register your toolset

To start the server:

uv run server.py --server_type=sse

🤖 How LLM Agents Use This

Agents connect to your MCP server and call tools directly using natural language instructions.

Example client setup:

from llama_index.tools.mcp import BasicMCPClient, McpToolSpec

mcp_client = BasicMCPClient("http://127.0.0.1:8000/sse")
mcp_tools = McpToolSpec(client=mcp_client)

This client will auto-discover tools like add_book, get_books, and more.

✨ Example instruction to LLM: “Add a book titled The Alchemist by Paulo Coelho, published in 1988, genre Fiction.”

The LLM will convert this into a SQL INSERT and call the add_book() tool.