Generate Markdown summaries for every TypeScript (.ts) and JavaScript (.js) file in any repository using an LLM (Ollama).
Great for quick codebase overviews, technical documentation, or onboarding material.
| Feature | Description |
|---|---|
| 🔍 Automatic repo scan | Walks a directory tree and collects all .ts and .js files. |
| 🤖 LLM-powered summaries | Uses Ollama (or any compatible LLM wrapper) to create concise, human-readable explanations of each file. |
| 🗂️ Markdown output | Saves per-file docs with original folder structure inside an output_docs/ directory. |
| 🏃 CLI entry point | Run with python -m docgen and pass --input / --output arguments. |
| ⏱️ Timing info | Prints how long each summary takes for quick performance insight. |
Requires Python 3.10+ and a running instance of Ollama with a model that understands code (e.g.,
codellama).
# 1. Install dependencies (Poetry)
poetry install
# 2. Activate virtualenv
poetry shell # or `poetry run ...` for one-offs
# 3. Run DocGen
python -m docgen --input /path/to/repo --output output_docsThis tool relies on Ollama for running local LLMs.
You must install Ollama and download a model mistral.
Follow instructions at https://ollama.com/download
scan_repo recursively finds .ts / .js files.
For each file:
Read the source code.
Send it to prompt_ollama for summarization.
write_markdown saves the summary as output_docs/<relative_path>.md.
CLI prints elapsed time per file so you can spot bottlenecks.
The core logic lives in src/docgen/main.py, so importing the module elsewhere (e.g., for tests) is straightforward: