repo-digest 0.1.0

Turn any repository into an AI-ready text bundle with safe defaults and rich analytics.

repo-digest

Turn any repository into an AI-ready text bundle with safe defaults and rich analytics.

Who is this for?

• Anyone who wants to paste a project into ChatGPT/Claude or create a quick, comprehensive repo digest.
• Works out-of-the-box on macOS, Linux, Windows.

Quickstart (60 seconds)

1. Install

   • pip install repo-digest
   • For precise token counts (optional): pip install "repo-digest[tiktoken]"

2. Export your repo

   • repo-digest . -o repo.txt

3. Preview first (optional)

   • repo-digest . -o repo.txt --preview

Safety first (defaults)

• Secrets are blocked by default (e.g., .env, secret, *.key, *.pem)
• Binary/large data files are excluded
• .gitignore respected by default
• To override secrets blocking (NOT recommended): --allow-secrets

Examples

• Export current repo: repo-digest . -o repo.txt
• Preview and check size: repo-digest . -o repo.txt --preview
• Enforce a size limit (bytes): repo-digest . -o repo.txt --max-bytes 5000000
• Ignore .gitignore: repo-digest . -o repo.txt --no-gitignore

Exit codes

• 0 success
• 1 runtime error (bad path, permission)
• 2 safety violation (secrets detected and not allowed)
• 3 exceeded size/limits

Troubleshooting

• Windows long paths: try running from a shorter path (e.g., C:\src)
• Encoding issues: files are read as UTF-8 with errors ignored
• Large repos: use --preview to estimate and --max-bytes to cap

FAQ

• Why are some files missing? They’re excluded by default to keep the export safe and useful. Use --no-gitignore or tweak locally if needed.
• Why token counts differ from my model? Install tiktoken for tokenizer parity; fallback uses an approximate word count.
• Can I include secrets? Not recommended. If you must: --allow-secrets (and understand the risk).

Roadmap (post-MVP)

• Markdown/JSON outputs, config file support
• GitHub URL input, chunking for huge repos
• Simple GUI if user demand is strong

License

• MIT