csvql MCP
csvql is a read-only, local MCP server that allows AI agents to query CSV files using SQL without importing data into a database. Built in Zig, it runs directly on the data, offering performance comparable to DuckDB with lower memory usage. It supports sandboxing via --root, limits output size to prevent context flooding, and provides audit logs for transparency.
Package and access
Local package with source-attributed installation metadata. Koskamo does not execute packages during review.
Checks and signals
Each line is separate evidence, not a trust score.
- Popularity 30 GitHub stars
What you can do with it
- Run SQL SELECT queries on local CSV files
- Aggregate data from large CSV datasets without loading into a database
- Enable AI agents to analyze CSV data in a sandboxed environment
Source and maintenance
- Last repository push
- Latest release
- v3.0.1
- Maintenance checked
Packages and installation
Source-provided package data. Koskamo has not executed or live-probed this package.
https://github.com/melihbirim/csvql/releases/download/v2.10.0/csvql-macos-aarch64.mcpb
{
"transport": "stdio",
"fileSha256": "a4d1cabbd098fd6c6ebbac272e51b36423e74d6eb2e4d20df72f6e1b4d2f25f2",
"identifier": "https://github.com/melihbirim/csvql/releases/download/v2.10.0/csvql-macos-aarch64.mcpb",
"registryType": "mcpb"
}Tools
Reported by the attributed source; not independently observed by Koskamo.
Maintenance history
Public check and source-refresh events. Editorial and ownership records remain private.
- · repository · passed · metadata changed · updated
- · backfill · recorded · Pre-expansion baseline; unobserved rich metadata remains unknown
Let your agent find its tools
Find servers like csvql MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.