Documentation

Everything you need to ship AI-ready context.

Learn how to use RepoContext to turn your GitHub repositories into accurate, structured documentation for AI coding agents.

Get up and running in 60 seconds.

Quick start

Paste any public GitHub repository URL into the homepage and click Analyze. Within seconds, RepoContext scans your repository and produces structured documentation that AI coding agents can use as reliable context.

text
github.com/owner/repository

The output includes a framework detection summary, dependency graph, build/test setup, and a quality score so you know how much you can trust the result.

Export once, use everywhere.

Supported formats

All plans include the canonical AGENTS.md export. Pro and Team unlock the rest:

  • AGENTS.md — the cross-tool standard for agent instructions.
  • CLAUDE.md — Claude-optimized prompt context.
  • Cursor Rules — .cursorrules for Cursor IDE.
  • Copilot Instructions — repository-level GitHub Copilot guidance.

Every claim is traceable.

Evidence panel

For each piece of generated documentation, RepoContext shows which files and config artifacts it was derived from. Click any statement to inspect the underlying evidence — no more guessing where a recommendation came from.

Already have a file? We'll grade it.

AGENTS.md audit

Upload an existing AGENTS.md and we'll audit it for accuracy, completeness, and staleness. You'll see a quality score plus a checklist of concrete fixes you can apply.

For teams & automation.

API reference

The same engine powers a JSON API. Send a repository URL and receive structured results you can pipe into your CI pipeline.

http
POST /api/analyze
Content-Type: application/json

{ "repoUrl": "github.com/owner/repository" }

Common questions.

FAQ

Is my code uploaded to your servers?

RepoContext only reads what it needs from the GitHub API. The full source tree is never written to disk on our side, and you can delete the cached metadata at any time from the dashboard.

Can I analyze private repositories?

Yes — Pro and Team plans support private repos via OAuth. Your code never leaves your control; we only read metadata, not file contents, by default.

How accurate is the output?

Every analysis ships with a quality score. Files with low confidence are flagged in the evidence panel so you know exactly where to double-check.