videodocgen 0.1.0

Convert demo videos into structured documentation, locally.

VideoDocGen

Convert demo/screen-recording videos into structured documentation — a Markdown "User Guide", extracted screenshots, and a structured document.json — using only local, pip-installable dependencies.

An open-source project by VOCSO Technologies Pvt. Ltd. · Maintained by Deepak Chauhan · github.com/vocso-com/videodocgen

Guarantees

• Local only. No cloud APIs, no API keys, no model downloads, no network calls at runtime.
• Deterministic & reproducible. Same input + config produce the same output. Runnable in CI.
• Extensible by design. OCR, understanding, rendering, and export are provider interfaces selected by config, so new implementations (cloud AI, additional OCR engines, more output formats) slot in without rework. See ARCHITECTURE.md.

Installation

VideoDocGen needs Python 3.12+ and two command-line tools on your PATH: FFmpeg (ffmpeg + ffprobe, for video decode and screenshot extraction) and Tesseract OCR (tesseract, for reading on-screen text).

1. Install the system tools

# macOS (Homebrew)
brew install ffmpeg tesseract

# Debian / Ubuntu
sudo apt-get update && sudo apt-get install -y ffmpeg tesseract-ocr

# Windows (winget)
winget install Gyan.FFmpeg UB-Mannheim.TesseractOCR

2. Install VideoDocGen

Not yet on PyPI — install from source. (pip install videodocgen is coming once the release workflow lands.)

git clone https://github.com/vocso-com/videodocgen.git
cd videodocgen
python3.12 -m venv .venv
source .venv/bin/activate          # Windows: .venv\Scripts\activate
pip install -e .

Use pip install -e ".[dev]" instead if you plan to run the tests or contribute (see Development).

3. Verify your setup

videodocgen doctor

This checks that FFmpeg, ffprobe, Tesseract, OpenCV, and PySceneDetect are all available and prints a ✓/✗ for each. Fix anything marked ✗ before running.

Quickstart

videodocgen run input.mp4

This runs the full pipeline against input.mp4 and writes a Markdown user guide, extracted screenshots, and document.json to the output directory (default docs/).

Other commands:

videodocgen batch DIR          # process every video in a directory
videodocgen config             # print the resolved effective configuration
videodocgen doctor             # verify FFmpeg/Tesseract/OpenCV are available

Run videodocgen doctor first if you're unsure whether your environment has the required system dependencies installed.

Development

pip install -e ".[dev]"
pre-commit install
ruff check .
black --check .
pytest

See CONTRIBUTING.md for the full workflow and ARCHITECTURE.md for the module layout.

About VOCSO

VideoDocGen is developed and maintained by VOCSO Technologies Pvt. Ltd., a software and design studio. We build it in the open so teams everywhere can turn product demos into documentation without manual screenshotting and writing.

GitHub org: @vocso-com · Maintainer: Deepak Chauhan. Contributions are welcome — see CONTRIBUTING.md.

License

MIT © VOCSO Technologies Pvt. Ltd.