Clips AI

Clips AI

Clips AI is an open-source Python library that helps developers automatically cut long narrative videos into short clips and reframe them to new aspect ratios.

Follow
Tags
ai-podcastlong-video-to-short-video-aiai-transcription
Last Updated Nov 10, 2026

About

Clips AI is an open-source Python library that converts longform video into short clips automatically. It addresses a common problem for teams working with lengthy spoken-word content: manually reviewing a full recording to find shareable segments and reframe them for vertical platforms is slow and repetitive work.

The library is designed for audio-centric, narrative-based videos such as podcasts, interviews, speeches, and sermons. Its clipping algorithm analyzes a video's transcript to identify segments worth extracting, and its resizing algorithm reframes the picture dynamically to keep the current speaker centered when converting between aspect ratios such as 16:9 to 9:16. Installation is a standard pip process, with transcription handled through the open-source WhisperX library and speaker tracking through Pyannote. The project is maintained by Clips AI, Inc., and its source code is available on GitHub.

Key Features

Automatic clip discovery: The ClipFinder component analyzes a video's transcript and returns a set of clips, each with explicit start and end times that developers can use for trimming.
Transcript-based segmentation: Because clips are located in the transcript rather than by visual analysis, the tool works best on speech-driven content like interviews, speeches, and sermons.
Transcription with WhisperX: Videos are transcribed through WhisperX, an open-source wrapper around Whisper that adds word-level start and stop timestamps needed for precise clipping.
Dynamic reframing: The resize function tracks the active speaker and reframes the video as it converts the aspect ratio, for example from 16:9 to 9:16 for vertical formats.
Minimal Python API: Core workflows such as transcribing, finding clips, and resizing take only a few lines of code, fitting naturally into existing Python pipelines.
Open source: The library can be inspected, modified, and self-hosted, and developers can integrate it into larger automated video workflows without per-seat licensing.

Pricing & Fees

  • Pricing: Clips AI is an open-source library, and no paid plans are advertised on the pages reviewed; check the official site for current information.
  • Installation requirements: Setup involves installing the clipsai pip package, WhisperX from its GitHub repository, plus the libmagic and ffmpeg system dependencies; a virtual environment is recommended.
  • Hugging Face token for resizing: Resizing requires a free Hugging Face access token because Pyannote is used for speaker diarization; the documentation states users are not charged for using Pyannote.
  • Best-fit content: The tool is intended for audio-centric, narrative-based videos rather than visually driven footage.

Who It's For

Clips AI is built primarily for developers who want programmatic control over video repurposing rather than a hosted editing app. Podcast producers can use it to turn full episodes into short highlights, and teams publishing interviews, speeches, or sermons can generate vertical clips for social platforms automatically. It also suits media startups and engineering teams that want to embed clip generation and aspect-ratio conversion into larger pipelines, since the entire workflow runs as ordinary Python code inside their own infrastructure.

FAQ

What is Clips AI?Clips AI is an open-source Python library that automatically converts longform, speech-driven videos into shorter clips and can reframe them into different aspect ratios such as 9:16.

Is Clips AI free to use?Yes, the library is open source and installed via pip; resizing requires a free Hugging Face access token for Pyannote, which the documentation says carries no charge.

Which dependencies does Clips AI need?It requires the clipsai package, WhisperX installed from its GitHub repository, plus libmagic and ffmpeg on the system, ideally inside a virtual environment.

How does Clips AI decide where clips start and end?It transcribes the video with WhisperX to obtain word-level timestamps, then analyzes that transcript to identify segments and returns each clip's start and end times.

What kinds of videos work best with Clips AI?Audio-centric, narrative-based content such as podcasts, interviews, speeches, and sermons works best, because clip detection relies on the spoken transcript.