Clips AI

Clips AI is an open-source Python library that automatically cuts long, narrative videos (podcasts, interviews, talks) into short clips and resizes them for any social platform.

It transcribes audio with WhisperX, finds semantic scene boundaries, splits the video into coherent segments, and re-crops each clip to the aspect ratio you need—portrait, square, or custom.

Speech-led, long-form content: podcasts, interviews, webinars, keynote speeches, sermons, lecture series—anything where the story is driven by dialogue.

Yes. Clips AI is a developer tool; you install it via pip and call its Python API to automate clipping and resizing.

WhisperX gives word-level timestamps; an improved TextTiling algorithm plus BERT embeddings spot topic shifts, creating natural cut points.

The core library is open source and free to use. You only pay for optional cloud compute (e.g., GPU transcription) if you choose not to run models locally.

Absolutely—one command outputs 9:16 for TikTok/Reels, 1:1 for Instagram feed, 4:5 for Facebook, or any custom resolution.

WhisperX (ASR), TextTiling + BERT (semantic segmentation), Pyannote (speaker diarization), PySceneDetect (scene cuts), MTCNN/MediaPipe (face detection).

pip install clipsai, add FFmpeg & libmagic, grab a WhisperX model, then follow the quick-start notebook to transcribe, clip, and resize your first video.