This Python tool accepts a video or audio file, transcribes it using OpenAI Whisper, embeds transcript segments using SentenceTransformers, and clusters them by topic using HDBSCAN.
- Automatic speech transcription (Whisper)
- Sentence segmentation
- Embedding via SentenceTransformers
- Clustering via HDBSCAN
- Topic summary generation (optional)
# Install deps
pip install -r requirements.txt
# Run segmentation
python segment_transcript.py --input examples/sample_video.mp4 --output segments.json[
{
"topic_id": 0,
"start": 12.3,
"end": 54.6,
"text": "Discussion about industry challenges..."
},
...
]