duck-mcp

lab1702/duck-mcp
★ 0 stars Python Data/Analytics Updated today
Read-only DuckDB MCP server: schema discovery and SQL queries over CSV, Parquet, JSON and Excel files, locally or on S3.
View on GitHub → 🔍 Audit Wallet Slippage →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "duck-mcp": {
      "command": "uvx",
      "args": [
        "duck-mcp"
      ]
    }
  }
}

Or install with pip: pip install duck-mcp

README Excerpt

An MCP server that lets an AI assistant explore and query data files with variants, globs, `https://` URLs and `s3://` buckets. It is **strictly read-only**: only a single `SELECT` (including `WITH`/`DESCRIBE`/`SUMMARIZE`/`SHOW`) or `EXPLAIN` statement runs per call. `INSERT`, `CREATE`, `COPY … TO`, `ATTACH`, `SET` and friends are rejected

Tools (6)

DUCKDB_MCP_MAX_BYTESDUCKDB_MCP_MAX_ROWSDUCKDB_MCP_MEMORY_LIMITDUCKDB_MCP_TIMEOUTpreview_filequery
📊
Working with market or time-series data?
The Polymarket Historical Dataset — 18.5M+ price snapshots across 18,600+ prediction markets (CSV + SQLite). Built for backtesting & ML.
View dataset →