Mache

v1.0.0Developer Toolsstable

IDE-style navigation for structured data — code, JSON, YAML. Jump to definitions, find callers, follow references. Available as an MCP server or a mounted folder.

agentic-researchart-ecosystemcapnprotocobracode-intelligence
Share:
41
Stars
0
Downloads
0
Weekly
0/5

What is Mache?

Mache is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to ide-style navigation for structured data — code, json, yaml. jump to definitions, find callers, follow references. available as an mcp server or a mounted folder.

IDE-style navigation for structured data — code, JSON, YAML. Jump to definitions, find callers, follow references. Available as an MCP server or a mounted folder.

This server falls under the Developer Tools category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • IDE-style navigation for structured data — code, JSON, YAML.

Use Cases

Jump to definitions and find references in code.
Navigate structured data (JSON, YAML) like an IDE.
Follow code references and understand architecture.
LicenseApache-2.0
Languagego
Versionv1.0.0
UpdatedMay 20, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx mache

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Mache

Mache is an IDE-style code intelligence engine that gives AI coding agents structured navigation capabilities over codebases — jump-to-definition, find-callers, cross-reference tracking, semantic search, and code smell detection — all exposed as MCP tools or a mountable FUSE filesystem. Built in Go and backed by a projected graph of your code, Mache lets agents explore large codebases the same way a developer would in an IDE rather than grepping through raw files. It supports 17 MCP tools covering function discovery, call chain navigation, write-back, schema inference, and community detection, making it especially effective for autonomous research and refactoring tasks.

Prerequisites

  • Go 1.21 or newer installed (to build from source)
  • Task (taskfile.dev) build tool installed: go install github.com/go-task/task/v3/cmd/task@latest
  • Claude Desktop with MCP support, or Claude Code CLI
  • Git to clone the repository
  • Optional: Docker for containerized deployment
1

Clone and build Mache

Clone the repository and build the mache binary using the Task build tool.

git clone https://github.com/agentic-research/mache.git
cd mache && task build && task install
2

Serve a codebase

Start the Mache server pointing at the source directory you want to analyze. By default it listens on port 7532.

mache serve /path/to/your/codebase
3

Connect Claude Code to Mache via HTTP transport

Register the running Mache server with Claude Code using the MCP HTTP transport.

claude mcp add --transport http mache http://localhost:7532/mcp
4

Configure Claude Desktop (alternative)

For Claude Desktop, you can run Mache in stdio mode. Add this to your claude_desktop_config.json, replacing the path with your actual codebase location.

{
  "mcpServers": {
    "mache": {
      "command": "mache",
      "args": ["serve", "--stdio", "/path/to/your/codebase"]
    }
  }
}
5

Optional: use Docker for containerized analysis

Build and run Mache as a Docker container, mounting your source code directory.

task image
docker load -i mache.tar
docker run --rm -i -v /path/to/codebase:/src mache:0.8.0 serve --stdio /src

Mache Examples

Client configuration

Claude Desktop config for Mache in stdio mode analyzing a local codebase.

{
  "mcpServers": {
    "mache": {
      "command": "mache",
      "args": ["serve", "--stdio", "/Users/yourname/projects/my-codebase"]
    }
  }
}

Prompts to try

Prompts that demonstrate Mache's code intelligence and navigation capabilities.

- "Find all callers of the authenticate() function in this codebase"
- "Show me the definition of the UserRepository class and its public methods"
- "What functions does the payment module call into?"
- "Find any dead code or unused functions in the src/utils directory"
- "Give me an overview of the codebase architecture and main modules"
- "Search for all places where the database connection is opened"

Troubleshooting Mache

mache binary not found after task install

The install task places the binary in your Go bin directory. Ensure $GOPATH/bin (typically ~/go/bin) is in your PATH. Run 'export PATH=$PATH:$(go env GOPATH)/bin' and add it to your shell profile.

Semantic search or type information tools return errors

The tools semantic_search, get_type_info, and get_diagnostics require ley-line-open enrichment. These are enriched tools that need additional setup. The 16 basic read-surface tools work without it.

Automatic binary download fails in CI or restricted environments

Set the environment variable MACHE_NO_LEYLINE=1 to disable automatic leyline binary discovery and download. This is recommended for CI pipelines and air-gapped environments.

Frequently Asked Questions about Mache

What is Mache?

Mache is a Model Context Protocol (MCP) server that ide-style navigation for structured data — code, json, yaml. jump to definitions, find callers, follow references. available as an mcp server or a mounted folder. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Mache?

Follow the installation instructions on the Mache GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Mache?

Mache works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Mache free to use?

Yes, Mache is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.

Browse More Developer Tools MCP Servers

Explore all developer tools servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "mache": { "command": "npx", "args": ["-y", "mache"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Mache?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides