SQLite Read-Only MCP Server
Enables safe, read-only SQL access to SQLite databases for AI agents, allowing schema exploration and SELECT queries with defense-in-depth protections.
README
MCP Server for SQLite (read-only) — safe SQL for your AI agent
A small, production-minded MCP server (Model Context Protocol) that gives an AI agent —
Claude Desktop, Claude Code, or any MCP client — safe, read-only SQL access to a SQLite
database. Point it at a .db file and the agent can explore schemas and run SELECT queries —
but never write, drop, or escape the database.
Built on the official MCP Python SDK (
FastMCP). The read-only safety core has zero MCP dependency and is unit-tested on its own — the server is a thin, auditable wrapper.
Why read-only matters
Handing an LLM a raw database connection is how you get a DROP TABLE in production. This server
enforces read-only at two independent layers:
- OS-level — the connection is opened with SQLite
?mode=ro+PRAGMA query_only, so any write fails at the engine. - Statement-level —
queryrejects anything that isn't a singleSELECT/WITH(no multi-statements, noINSERT/UPDATE/DELETE/DROP/PRAGMAwrites), and caps row counts.
Defense in depth: even if one layer is bypassed, the other still holds. (There's a test that proves it.)
What the agent gets (all 3 MCP primitives)
| Primitive | Name | Does |
|---|---|---|
| 🔧 tool | list_tables |
List the tables in the database |
| 🔧 tool | describe_table |
Column schema (name, type, nullability, PK) |
| 🔧 tool | query |
Run a single read-only SELECT/WITH, capped at max_rows |
| 📄 resource | schema://tables |
The whole DB schema as text |
| 💬 prompt | explore_database |
A ready-made "explore this DB" prompt |
Quick start (≈ 1 minute)
git clone https://github.com/skycandykey1/mcp-sqlite-server
cd mcp-sqlite-server
pip install -r requirements.txt
# make a sample DB to play with
python examples/make_sample_db.py
# inspect it live in the MCP dev inspector
SQLITE_DB_PATH=examples/sample.db mcp dev mcp_sqlite/server.py
# ...or run the server directly
SQLITE_DB_PATH=examples/sample.db python -m mcp_sqlite.server
Use it from Claude Desktop
Add this to your Claude Desktop config (Settings → Developer → Edit Config), using absolute paths —
see examples/claude_desktop_config.example.json:
{
"mcpServers": {
"sqlite-readonly": {
"command": "python",
"args": ["-m", "mcp_sqlite.server"],
"cwd": "/absolute/path/to/mcp-sqlite-server",
"env": { "SQLITE_DB_PATH": "/absolute/path/to/your.db" }
}
}
}
Restart Claude Desktop, then ask: "What tables are in my database? Show me the top 5 orders by amount."
How it fits together
MCP client (Claude Desktop / Claude Code)
│ MCP over stdio
▼
mcp_sqlite/server.py ← thin FastMCP wrapper (tools / resource / prompt)
│
▼
mcp_sqlite/db.py ← read-only core (no MCP dep, fully unit-tested)
│ ?mode=ro + query_only + SELECT-only guard
▼
your .db (read-only)
Run the tests
pip install -r requirements-dev.txt
python -m pytest -q # offline — no MCP client, no API key needed
License
MIT — see LICENSE.
💼 Built by
skycandykey1— available for AI agent & automation contract work. I build agents, MCP servers, and LLM automation. →skycandykey456@gmail.com·https://github.com/skycandykey1
<!-- After publishing (SEO): About: "Read-only SQLite MCP server (Model Context Protocol) — safe SQL access for AI agents" Topics (≥5): mcp, mcp-server, model-context-protocol, sqlite, claude, anthropic, ai-agent, python Add a social preview image. Pin this repo on your profile, next to ai-agent-starter. -->
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.