PluginBench
MCP Server
Active
MIT

io.github.Dave-London/pare-search MCP Server

io.github.Dave-London/pare-search

Structured code search with ripgrep and fd—typed JSON output for reliable agent workflows.

What is the io.github.Dave-London/pare-search MCP server?

Pare Search is an MCP server that wraps ripgrep and fd to provide structured, schema-validated JSON output for code search operations. Instead of parsing fragile terminal text, agents receive clean typed data with consistent field names across platforms and tool versions.

Pare Search exposes ripgrep and fd as MCP tools that return JSON instead of raw terminal output. This eliminates parsing errors and reduces token usage by returning only the data agents need. Use it to search code, find files, and filter results reliably without brittle regex parsing.

How to install io.github.Dave-London/pare-search

Copy-paste configuration for popular MCP clients.

transport: stdio
Config generated by PluginBench — verify against the source before use.
~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "pare-search": {
      "command": "npx",
      "args": [
        "-y",
        "@paretools/search"
      ]
    }
  }
}

Tools & capabilities

Tools this server exposes to the agent.

  • ripgrep search — Search code with ripgrep, returning matches as structured JSON with file, line, column, and match text.
  • fd find — Find files using fd with pattern matching, returning results as a structured list with paths and metadata.

Use cases

  • Search a codebase for specific patterns or function names and get results as typed JSON
  • Find all files matching a pattern (e.g., all test files) without parsing terminal output
  • Filter search results by file type, size, or other criteria using fd options
  • Locate references to a variable or function across a project reliably
  • Build agent workflows that depend on accurate file discovery and code search results

io.github.Dave-London/pare-search MCP server FAQ

What is Pare Search?

Pare Search wraps ripgrep (rg) and fd as MCP tools that return schema-validated JSON instead of raw terminal text. Agents get typed, structured results they can act on directly without custom parsing.

How do I install it?

Install via npm: `npm install @paretools/search` or use `npx @paretools/search`. For Claude Code or Cursor, use the setup guide at https://github.com/Dave-London/Pare/docs/setup/.

Is it free?

Yes, Pare is open-source under the MIT license. It wraps free, open-source tools (ripgrep and fd) and requires no external API keys or authentication.

What are the token savings?

Pare returns structured JSON that is typically 65–95% smaller than raw CLI output. For example, a verbose git log might use 4,992 tokens raw but only 382 tokens structured—a 92% reduction.

Can I limit which tools are available?

Yes, use environment variables like `PARE_SEARCH_TOOLS=ripgrep,fd` to filter tools, or `PARE_TOOLS=search:ripgrep` to restrict across all servers.

What Node.js version is required?

Pare requires Node.js >= 20. Check your version with `node --version` and upgrade if needed.

README (reference)

Source of truth, from the repository.

<h1><img src="assets/logo.png" alt="" width="80" valign="middle" />&nbsp;&nbsp;Pare</h1>

CI codecov npm Downloads TypeScript License: MIT Node.js >= 20 OpenSSF Scorecard OpenSSF Best Practices

Reliable, structured CLI output for AI agents — no more parsing fragile terminal text.

Pare provides MCP servers that wrap common developer tools (git, npm, docker, test runners, etc.) and return clean, schema-validated JSON instead of raw terminal text. Agents get typed data they can act on directly, without brittle string parsing.

The Problem

Parsing CLI output is fragile. Raw terminal text includes ANSI escape codes, decorative headers, progress bars, locale-specific formatting, and platform differences that break agent workflows in subtle ways. An agent that works fine with git status on macOS may fail on Windows because the output format changed. A test runner's summary line might shift between versions, silently breaking a regex.

Pare eliminates this entire class of errors by returning schema-validated JSON with consistent field names, regardless of platform, tool version, or locale. As a bonus, structured output is significantly smaller — agents use fewer tokens per tool call:

Tool CommandRaw TokensPare TokensReduction
docker build (multi-stage, 11 steps)3732095%
git log --stat (5 commits, verbose)4,99238292%
npm install (487 packages, warnings)2414183%
vitest run (28 tests, all pass)1963980%
cargo build (2 errors, help text)43613868%
pip install (9 packages, progress bars)28810165%
cargo test (12 tests, 2 failures)35119046%
npm audit (4 vulnerabilities)28718536%

Token estimates use ~4 chars/token. The biggest savings appear on verbose commands (builds, installs, tests). For simpler tools like eslint or tsc, the main advantage is reliable structured data — agents can use typed JSON directly rather than parsing strings.

How It Works

Each Pare tool returns two outputs:

  • content — human-readable text, for MCP clients that display it
  • structuredContent — typed, schema-validated JSON, ready for agents to process

This uses MCP's structuredContent and outputSchema features to provide type-safe, validated data that agents can rely on without custom parsing.

Example: git status

Raw git output (~118 tokens):

On branch main
Your branch is ahead of 'origin/main' by 2 commits.
  (use "git push" to publish your local commits)

Changes to be committed:
  (use "git restore --staged <file>..." to unstage)
        modified:   src/index.ts
        new file:   src/utils.ts

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
        modified:   README.md

Untracked files:
  (use "git add <file>..." to include in what will be committed)
        temp.log

Pare structured output (~59 tokens):

{
  "branch": "main",
  "upstream": "origin/main",
  "ahead": 2,
  "staged": [
    { "file": "src/index.ts", "status": "modified" },
    { "file": "src/utils.ts", "status": "added" }
  ],
  "modified": ["README.md"],
  "deleted": [],
  "untracked": ["temp.log"],
  "conflicts": [],
  "clean": false
}

50% fewer tokens. Zero information lost. Fully typed. Savings scale with output verbosity — test runners and build logs see 80–92% reduction.

Available Servers (28 packages, 240 tools)

Install only the servers relevant to your stack — most projects need just 2–4. The full catalog covers a wide range of ecosystems so Pare works wherever you do.

CategoryServersToolsWraps
Version Controlgit, github55git, gh
Languages & Packagesnpm, python, cargo, go, deno, bun, nix, dotnet, ruby, swift, jvm101npm, pip, cargo, go, deno, bun, nix, dotnet, gem, swift, gradle, maven
Build, Lint & Testbuild, lint, test, cmake, bazel23tsc, esbuild, vite, webpack, eslint, prettier, biome, vitest, pytest, jest
Infrastructuredocker, k8s, infra, security, remote40docker, kubectl, helm, terraform, ansible, trivy, ssh
Utilitiessearch, http, make, process, db21ripgrep, fd, curl, make, just, psql, mysql, redis, mongosh

Tool Schemas — detailed response examples and field descriptions for every tool. See also Tool Response Examples for quick JSON samples.

Quick Setup

# 1. Configure MCP servers (non-interactive)
npx @paretools/init --client claude-code --preset web

# 2. Add agent rules to your project
#    (append to existing CLAUDE.md, or copy if new)
cat node_modules/@paretools/init/rules/CLAUDE.md >> CLAUDE.md

# 3. Restart your client session

# 4. Validate
npx @paretools/init doctor

Available presets: web, python, rust, go, jvm, dotnet, ruby, swift, mobile, devops, full

Setup Guides by Client

Claude CodeClaude DesktopCursor
VS Code / CopilotWindsurfCline / Roo Code
OpenAI CodexGemini CLIZed
Continue.dev

Full Quickstart Guide — presets, ecosystem mapping, validation

Manual Configuration — config paths and formats for all clients

Agent Integration Guide — rule files, hooks, CLI-to-MCP mapping

Configuration

Tool Selection

By default, every Pare server registers all of its tools. If a server exposes tools you don't need — or you want to limit which tools are available to an agent — you can filter them with environment variables.

Per-server filter — restrict a single server's tools:

# Only register status and log in the git server
PARE_GIT_TOOLS=status,log npx @paretools/git

Universal filter — restrict tools across all servers:

# Only register these specific tools across any server
PARE_TOOLS=git:status,git:log,npm:install npx @paretools/git

Disable all tools — set the env var to an empty string:

PARE_GIT_TOOLS= npx @paretools/git   # no tools registered
Env VarScopeFormatExample
PARE_TOOLSAll serversserver:tool,...git:status,npm:install
PARE_{SERVER}_TOOLSOne servertool,...status,log,diff

Rules:

  • No env var = all tools enabled (default)
  • PARE_TOOLS (universal) takes precedence over per-server vars
  • Server names use uppercase with hyphens replaced by underscores (e.g., PARE_MY_SERVER_TOOLS)
  • Whitespace around commas is ignored

Common patterns:

# Read-only git (no push, commit, add, checkout)
PARE_GIT_TOOLS=status,log,diff,branch,show

# Minimal npm
PARE_NPM_TOOLS=install,test,run

# Only specific tools across all servers
PARE_TOOLS=git:status,git:diff,npm:install,test:run

In JSON MCP config, set via the env key:

{
  "mcpServers": {
    "pare-git": {
      "command": "npx",
      "args": ["-y", "@paretools/git"],
      "env": {
        "PARE_GIT_TOOLS": "status,log,diff,show"
      }
    }
  }
}

Troubleshooting

IssueSolution
npx not found / ENOENT on WindowsUse cmd /c npx wrapper (see your client's setup guide)
Slow first startRun npx -y @paretools/git once to cache, or install globally: npm i -g @paretools/git
Node.js version errorPare requires Node.js >= 20
NVM/fnm PATH issuesUse absolute path to npx: e.g., ~/.nvm/versions/node/v22/bin/npx
MCP connection timeoutSet MCP_TIMEOUT=30000 for Claude Code, or increase initTimeout in client config
Too many tools filling contextUse tool selection env vars to limit tools, or only install the servers you need

Contributing

Each server is a self-contained package. See CONTRIBUTING.md for the full guide.

License

MIT

Related MCP servers

Auto-detects test frameworks and returns typed JSON—pytest, jest, vitest, mocha.

130
TypeScript
MIT
View repository →

Structured Python tool output (ruff, mypy, pip-audit, pip-install) as typed JSON for AI agents.

130
TypeScript
MIT
View repository →

Code search with ripgrep and fd, returning structured JSON for token-efficient AI agent workflows.

130
TypeScript
MIT
View repository →

Auto-detects test frameworks (pytest, jest, vitest) and returns structured, schema-validated JSON results instead of raw terminal output.

130
TypeScript
MIT
View repository →

A Model Context Protocol (MCP) server for Uptime Kuma version 2.

29
TypeScript
MIT
View repository →

MCP for AI agents: financing, skills, lending on Base

View repository →