title	NeuronClip Backend
emoji	🚀
colorFrom	blue
colorTo	blue
sdk	gradio
sdk_version	6.20.0
python_version	3.12
app_file	app.py
pinned	false
license	mit

ClipNeuron

Free, local-first AI clipping pipeline for long-form podcasts.

MVP-1 (current)

• Extract audio with ffmpeg
• Transcribe with faster-whisper (open-source models)
• Segment transcript into candidate moments
• Score candidates with heuristic virality signals
• Output clips.json with ranked candidates

Requirements

• Python 3.10+
• ffmpeg installed and available in PATH
• Free open-source models (faster-whisper downloads them automatically)

Install

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Run (MVP-1)

python -m clipneuron.cli \
  --input path/to/podcast.mp4 \
  --output out/clips.json

Render top clips (MVP-2)

python -m clipneuron.cli \
  --input path/to/podcast.mp4 \
  --output out/clips.json \
  --render \
  --top 5 \
  --clips-dir out/clips

Notes

• This pipeline uses only free resources and open-source models.
• MVP-2 adds clip rendering, 9:16 cropping, and basic burned-in captions.