Search and filter code repositories using ripgrep and fd with structured JSON queries. Built for developers who need programmatic, type-safe code search integrated into AI workflows.
io.github.Dave-London/pare-search
Local install
STDIO
No auth required
How models use it and what it is built for.
Search and filter code repositories using ripgrep and fd with structured JSON queries. Built for developers who need programmatic, type-safe code search integrated into AI workflows.
Local install — runs as a subprocess.
Where to find authoritative docs and source for Dave London Pare Search.
Paste any of these into Agent Studio after connecting Dave London Pare Search.
Common questions about connecting and running Dave London Pare Search.
What search tools does Pare Search use?
Pare Search wraps ripgrep (for fast text search) and fd (for file discovery) and exposes them as typed JSON operations. This lets you perform structured code searches programmatically instead of via CLI.
How do I install Pare Search?
Run `npx @paretools/search@0.8.1` to install the MCP server. It communicates via stdio transport, so it integrates directly with MCP-compatible clients.
Can I use Pare Search with my existing ripgrep workflows?
Yes—Pare Search is built on ripgrep and fd, so it supports their core flags and patterns. The main difference is results come back as typed JSON instead of raw CLI output, making them easier to parse and chain in code.
What format do search queries use?
Pare Search accepts structured JSON queries for both ripgrep (text search) and fd (file matching). Check the server documentation for the exact schema and supported options.
Is Pare Search free?
Pare Search is open-source and available via npm. There are no licensing fees; you only need Node.js and an MCP-compatible client to run it.
MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 30+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.
Open Agent Studio