Repomix

{
  "service_type": "platform",
  "base_url": "https://repomix.com",
  "auth_method": "none",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "free",
    "details": {
      "notes": "Repomix is an open-source CLI tool and library with no paid tiers. The website acts as documentation and a web interface for the CLI. There is no commercial pricing model."
    }
  },
  "rate_limits": {},
  "capabilities": [
    "CLI tool to pack repositories into a single AI-friendly file",
    "Supports local and remote repositories (GitHub, branches, tags, commits)",
    "Output formats: XML, Markdown, JSON, plain text",
    "Tree-sitter based code compression to reduce token usage",
    "Comment removal from code output",
    "Token counting and split output for large repos",
    "Secret detection via Secretlint safety checks",
    "Configurable include/ignore patterns via JSON/JSONC/JSON5/JS/TS config",
    "MCP (Model Context Protocol) server mode for AI assistants",
    "Agent Skills generation (Claude Agent Skills, Repomix Explorer Skill)",
    "Claude Code plugins with slash commands",
    "GitHub Actions integration",
    "Watch mode for automatic re-packing on file changes",
    "Usable as a Node.js library",
    "Browser extension and VS Code extension integrations",
    "Git diff and commit log packing",
    "Custom instructions injection into output"
  ],
  "raw_analysis": "Repomix is an open-source developer tool, not a hosted SaaS platform with a public REST API. It is distributed primarily as a Node.js CLI (installable via npx, npm, Yarn, Bun, Homebrew, Docker) and as a library, and its website (repomix.com) serves as documentation, a privacy policy, and a web-based entry point. There is no public API server, no authentication, and no pricing model — usage is free and open source.\n\nCore purpose: pack a local or remote code repository into a single AI-friendly file (XML, Markdown, JSON, or plain text) to feed LLMs such as Claude, ChatGPT, Gemini, Grok, DeepSeek, Perplexity, and MCP-compatible agents. Key differentiators include Tree-sitter-based code compression, comment removal, token counting, Secretlint-based secret detection, and MCP server mode allowing AI assistants to pack/search/read codebases directly.\n\nIntegrations and ecosystem: MCP server, Claude Code plugins, Claude Agent Skills, GitHub Actions, VS Code extension, browser extension, and a community projects ecosystem. It also exposes functionality as a Node.js library for programmatic use.\n\nMaturity: well-established open-source project with comprehensive documentation (guide pages for installation, usage, configuration, output formats, security, watch mode, remote repo processing, etc.). Active ecosystem of community extensions and integrations. Target audience is developers and AI practitioners who need to provide codebase context to LLMs, perform AI-assisted code review, onboarding, refactoring, security audits, and documentation generation.\n\nSince there is no public REST API, the endpoints array is empty and auth is 'none'. Any automation would go through the CLI, the MCP server protocol, or the Node.js library rather than HTTP endpoints."
}

3/3 tests passed

{
  "overall": 86,
  "dimensions": {
    "token_efficiency": 10.0,
    "first_try_success": 9.0,
    "response_parseability": 9.0,
    "error_clarity": 8.0,
    "doc_quality": 9.0,
    "auth_simplicity": 10.0,
    "latency": 10.0,
    "consistency": 8.5
  },
  "pricing_normalized": {
    "model": "free",
    "open_source": true,
    "notes": "No paid tiers; fully open-source CLI/library. Zero-cost adoption."
  },
  "issues": [
    "No security headers on responses (CSP, HSTS, etc.) — minor trust signal gap",
    "latency baseline is excellent but no status page/uptime history referenced",
    "No explicit latency SLA documented for the hosted web interface"
  ],
  "recommendations": [
    "Highlight the llms.txt and structured doc endpoint as a first-class agent-onboarding surface",
    "Surface MCP server mode and Agent Skills generation more prominently — key differentiators for AI workflows",
    "Add security headers to improve trust signals for enterprise/agent deployments",
    "Publish a status page or uptime history to strengthen consistency perception"
  ]
}