SQLite MCP Server

What it does

• Read/write SQL over local .db files: read_query, write_query, create_table, list_tables, describe-table
• Bonus BI demo: append_insight tool + memo://insights resource that aggregates findings
• Python server (`mcp-server-sqlite`); point it at a file with --db-path

When to use it

Fastest way to let Claude explore a local dataset — CSV-to-sqlite a file and ask questions. The insight memo makes it a decent self-contained data-analysis demo.

Getting started

Add to .mcp.json under mcpServers.sqlite:

{
  "command": "uvx",
  "args": ["mcp-server-sqlite", "--db-path", "~/test.db"]
}

Docker: `{"command": "docker", "args": ["run", "-i", "--rm", "-v", "mcp-sqlite:/mcp", "mcp/sqlite", "--db-path", "/mcp/test.db"]}`.

Things to know

Archived reference implementation (servers-archived) — no updates. Unlike the Postgres server it is NOT read-only: write_query and create_table mutate your file, so point it at a copy, not your only database.