Skip to content

Open-source AI podcast clipper. Generate vertical clips with face tracking and burned-in captions. CLI, MCP server, web app.

934 B
Raw Blame
# Hermetic native-CLI runtime deps. Transcription is whisper.cpp (a native
# binary), so openai-whisper/torch are intentionally absent. That is the ~2GB
# the native install saves. --engine whisper-py remains a dev/source-only option
# via backend/requirements.txt.
# Direct dependencies are pinned to the versions the previous ranges resolved
# to on 2026-11-03. Transitive dependencies still float. Bump deliberately.
opencv-python-headless==5.0.0.93
numpy==2.5.4
# Audio-event detection (YAMNet laughter/reaction channel) runs on ONNX Runtime,
# no torch/TF, so it stays on the hermetic native path.
onnxruntime==1.31.1
# --engine omnilingual: Meta's Omnilingual ASR CTC model via sherpa-onnx, for
# languages whisper.cpp handles poorly. No torch, stays on the hermetic path.
sherpa-onnx==2.0.1
Pillow==13.0.0
questionary==2.1.1
python-dotenv==1.2.4
yt-dlp==2026.10.26
google-api-python-client==2.201.0
google-auth-oauthlib==1.5.0