An AI video editing agent that turns natural-language instructions into clip search, script writing, music and font selection, and edits.
blixvip/easyedit
A local-first Python tool that turns a movie title into a fan edit: a captioned best-speech clip followed by a montage cut to a song's beat.
- Stars
- 162
- Forks
- 13
- Open issues
- 0
- Last commit
- 12 days ago
What is it?
easyedit takes a movie name and produces a finished 1920×1080, 60fps fan edit. An LLM picks the film's most iconic speech and a fitting song, then yt-dlp downloads the scene, montage footage and music. faster-whisper provides word-level timestamps for the captions, and OpenCV YuNet tracks the speaker's face so the camera can follow them.
The montage is built by detecting scene cuts in the footage, scoring each shot on motion, sharpness, contrast and faces, and beat-tracking the song so cuts land on real beats and the montage starts on the drop. The visual effects are written as a HyperFrames composition, covering animated captions, whip transitions, zoom pulses and a black-and-white closing shot.
Every stage caches its output in a per-movie job folder, and you can pin or exclude shots with a curate.json file, edit the quote, or re-render. It works with an existing Claude Code or Codex login and falls back to heuristics without one. A web UI and a local API are also included.
Key features
- 01Generates a captioned speech clip and a beat-synced montage from a movie title alone.
- 02Transcribes every word with faster-whisper, using GPU when available.
- 03Tracks the speaker's face so the camera can push in and follow.
- 04Scores and filters montage shots, removing black bars and near-duplicates.
- 05Caches every stage and lets you pin, exclude or hand-edit shots and the quote.
- 06Includes a local web UI and an API that AI agents can drive.
Use cases
- Making a movie fan edit for social media without manual video editing.
- Experimenting with beat-synced montage and word-level caption effects.
- Letting an AI agent produce a finished video from a single prompt.
Quick start
- 1
Clone the repository
git clone https://github.com/blixvip/easyedit && cd easyedit - 2
Install Node dependencies
npm install - 3
Install Python dependencies
pip install -r requirements.txt - 4
Generate a fan edit for a movie
python -m easyedit "The Wolf of Wall Street"
Things to know
- Requires FFmpeg on PATH, Node.js 22+ and Python 3.10+ installed first.
- Rendering is slow, about 8 minutes for a 40-second edit on a GTX 1650.
- Downloads YouTube footage via yt-dlp, which can fail on age-restricted or challenged clips.
- Works best with an AI account (Claude Code or Codex) for scene and song picks; otherwise it falls back to heuristics.
Topics
FAQ
- What is easyedit?
- A local-first Python tool that turns a movie title into a fan edit: a captioned best-speech clip followed by a montage cut to a song's beat. easyedit takes a movie name and produces a finished 1920×1080, 60fps fan edit. An LLM picks the film's most iconic speech and a fitting song, then yt-dlp downloads the scene, montage footage and music. faster-whisper provides word-level timestamps for the captions, and OpenCV YuNet tracks the speaker's face so the camera can follow them.
- Is easyedit free to use? What license does it have?
- easyedit is open source under the MIT license. Check the license terms before using it in commercial products.
- Is easyedit still maintained?
- Actively maintained. The last commit to easyedit was on Oct 8, 2026.
- What can I use easyedit for?
- Making a movie fan edit for social media without manual video editing. Experimenting with beat-synced montage and word-level caption effects. Letting an AI agent produce a finished video from a single prompt.
- How do I install easyedit?
- Clone the repository: git clone https://github.com/blixvip/easyedit && cd easyedit Install Node dependencies: npm install Install Python dependencies: pip install -r requirements.txt Generate a fan edit for a movie: python -m easyedit "The Wolf of Wall Street"
- Where can I get easyedit?
- The source code and installation instructions are on GitHub: https://github.com/blixvip/easyedit
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An Agent Skill and MCP server giving coding agents like Claude Code, Cursor and Codex 42 local FFmpeg video and audio editing tools.
A Claude Code skill that downloads YouTube videos, splits them into AI-generated chapters, clips segments, and burns bilingual subtitles into the output.
A Python command-line tool that records TikTok live streams, including audio and video, in manual, automatic, or followers mode.
A browser-based video subtitle workstation that edits, styles and burns subtitles into local MP4 files, with optional on-device Whisper transcription.
Desktop app that transcribes a video with AI and burns styled captions directly into the output video.