Purpose
RepoCrunch is used to analyze any public GitHub repository into structured JSON. According to official information, it does not use AI or large language models, but instead runs fully deterministically: given a repository, it returns the tech stack, dependencies, architecture, health metrics, and security signals, with output as clean, consistent JSON.
Main Capabilities
- Tech stack and dependency analysis: identifies the tech stack and dependencies used by the repository.
- Architecture and health metrics: outputs architecture information and repository health metrics.
- Security signals: provides security-related signals (note: in the information, "Secrets regex scanning" is listed under "Not yet implemented, but planned").
- Framework detection: covers more than 40 frameworks across the supported ecosystems, including FastAPI, Django, React, Next.js, Spring Boot, Rails, Gin, Actix, and others, as well as more frameworks not listed individually.
Usage
RepoCrunch can be used as:
- A Python library
- A CLI tool
- A REST API
- An MCP server (the information mentions it is suitable for Claude, Cursor, etc.)
Runtime requirement: Python 3.11+. The information shows it can also be used with uv.
Installation:
- Install from PyPI:
pip install repocrunch - Install from source (the information gives the heading "Install from source")
- Install only what you need (the information gives the heading "Install just what you need")
GitHub Token (optional): without setting a token, 60 API calls per hour are allowed; with it, 5,000 per hour. The token also allows analysis of private repositories you have permission to access.
MCP server: the usage description given in the information is "for Claude, Cursor, etc.", that is, for use by Claude, Cursor, etc.
Released and Planned
Released: RepoCrunch has been released on PyPI and can be installed via pip install repocrunch .
Features planned but not yet implemented (the information explicitly marks them as "Not yet implemented, but planned"):
- Secrets regex scanning - detect leaked API keys, tokens, and passwords in the file tree
- Architecture type classification - distinguish libraries, applications, and frameworks
- API rate limiting - provide per-key throttling for the REST API
- npm/npx package -
npx repocrunch analyze owner/repo - Vulnerability scanning - detect known CVEs in dependencies
- Comparison mode - perform side-by-side analysis of multiple repositories
- Historical tracking - track changes in repository health over time
- Platform deployment - Apify Store, Smithery, mcp.so, RapidAPI
Suitable Scenarios
Suitable for users who need deterministic, repeatable structured analysis of GitHub repositories and want to obtain the tech stack, dependencies, architecture, health metrics, and security signals in JSON form; also suitable for integrating repository analysis into Python programs, command-line workflows, REST API services, or MCP workflows (such as Claude, Cursor).
Limitations
- The analysis target is GitHub repositories (the information says "any public GitHub repository"); analyzing private repositories requires a token and access permission.
- It does not use AI or LLMs and is deterministic analysis, so it does not provide model-based inference capabilities.
- Without a token, the GitHub API limit is 60 calls per hour; with a token, it is 5,000 per hour.
- The runtime environment requires Python 3.11+.
- The features listed in the information, such as Secrets scanning, architecture type classification, API rate limiting, npm/npx package, vulnerability scanning, comparison mode, historical tracking, and platform deployment, are currently not implemented and are only planned.
