gdc-vtt-capture

GDC-vtt-capture is a Python script that helps you download captions from a GDC talk video. It will fetch and merge segmented GDC caption .vtt chunks into a single text file. You can use other AI tools to summarize the talk by using this captions file.

How To Use

1. Login to your GDC Vault account and open the target session page.
2. Play the video, then open browser DevTools -> Network.
3. Find any .vtt request and copy its URL.
4. Run this script with that URL as the input parameter.

python3 capture_gdc_vtt.py "<vtt_segment_url>"

An example can be found in the repo, check run_vtt_capture_example.sh file.

How It Works

• Any single .vtt URL is used as a sample to get the common caption path template.
• The script brute-forces chunk ids in a while-loop (..._%d.vtt) to fetch the full stream.
• If some chunk fails (e.g., timeout or missing file), it logs and skips that chunk.
• After more than --max-404 404 responses, it assumes there are no more caption chunks and writes the merged output file.

Environment Setup

The tools is lightweighted, what you need are

• Python 3
• requests

License

MIT. See LICENSE.