Last updated 08-15-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Clips AI
Clips AI is an open-source Python library that turns long-form videos into short clips and reframes them for vertical feeds. You install it with pip, transcribe a podcast or interview with WhisperX, then call ClipFinder to segment the transcript into clip start and end times. Developers wire those timestamps into their own pipelines instead of uploading files to a hosted clip editor.
Consumer clip tools hide the segmentation logic behind a web upload form. Clips AI exposes TextTiling with BERT embeddings on Whisper transcripts, plus a separate resize path that tracks the active speaker with Pyannote diarization, PySceneDetect, and face detection. That stack targets narrative audio like podcasts and sermons, not quick meme cuts from silent b-roll.
Engineers building podcast repurposing workflows, media startups, and research teams who need programmatic clip boundaries and 16:9 to 9:16 reframing on their own servers. It fits teams that already run Python and want Hugging Face Pyannote tokens in their resize step rather than a black-box SaaS export button.
pip install clipsai plus WhisperX from GitHub for word-level transcription timestamps
ClipFinder segments transcripts with TextTiling and BERT embeddings to return start and end times
Resize function reframes 16:9 video to 9:16 using Pyannote speaker diarization
Face detection combines MTCNN and MediaPipe with PySceneDetect scene changes
MediaEditor trims or resizes source files to exported clip paths in Python
Designed for podcasts, interviews, speeches, and sermons with narrative audio
WhisperX batch_size defaults to 16 with ISO 639-1 language autodetection support
Open-source library you can run on your own infrastructure
Word-level WhisperX timestamps feed precise clip boundaries
Speaker-aware resize pipeline reframes to 9:16 automatically
Clear Python API for transcribe, clip, trim, and resize steps
Built specifically for narrative long-form audio content
Requires Python setup plus WhisperX and Pyannote dependencies
Pyannote Hugging Face token required for resize features
No hosted GUI for non-developer users on the site
Narrative audio focus means it is not aimed at silent action footage
Is Clips AI free to use?
Yes, Clips AI is an open-source Python library distributed on clipsai.com. You install it with pip and run clipping and resizing code locally without a paid subscription listed on the documentation site.
What video types does Clips AI support?
Clips AI targets audio-centric, narrative videos such as podcasts, interviews, speeches, and sermons. The homepage documentation says the clipping algorithm analyzes transcripts rather than action-heavy silent footage.
How does Clips AI find clip boundaries?
Clips AI transcribes with WhisperX, then ClipFinder applies TextTiling with BERT embeddings on the transcript. The clip reference page says it detects topic shifts at sentence granularity to return start and end times.
How does Clips AI resize video to vertical format?
The Clips AI resize function uses Pyannote speaker diarization, PySceneDetect scene changes, and MTCNN plus MediaPipe face detection. Clips AI returns crop segments that follow whoever is speaking while converting aspect ratio to 9:16 by default.
Do I need a Hugging Face token for Clips AI?
Yes, for resizing with Clips AI. The resize documentation requires a Pyannote authentication token from Hugging Face. Clips AI docs note you are not charged for using Pyannote through that token.
Can Clips AI transcribe non-English audio?
Yes. The Clips AI Transcriber class accepts an optional ISO 639-1 language code and defaults to autodetecting the media language when none is provided. WhisperX handles the underlying transcription for Clips AI.
Does Clips AI have a hosted web app?
No. Clips AI is a developer library with Python API examples on clipsai.com. You run transcription, clipping, and resizing in your own environment rather than through a browser upload interface.