Analysis updated 2026-05-18
Trim, merge, and color grade raw footage by describing the edit in plain English.
Auto generate subtitled vertical shorts from a longer talking head video.
Transcribe a video locally and burn the subtitles back into the footage.
Export a finished cut with presets sized for YouTube, Instagram, or TikTok.
| akmessi/vex | gcarq/inoxunpack | liumengxuan04/shushu-internship-tool | |
|---|---|---|---|
| Stars | 36 | 36 | 36 |
| Language | Python | Python | Python |
| Last pushed | — | 2018-08-04 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | moderate | easy | easy |
| Complexity | 3/5 | 2/5 | 2/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Python 3.11+ and FFmpeg on PATH, Node.js needed only for advanced generated visuals.
Vex is an open source tool that lets you edit video by typing plain English instructions into a terminal instead of using a traditional video editor. You launch it, point it at a video file, and tell it what you want done. Behind the scenes it uses FFmpeg, MoviePy, and a language model that plans and runs the actual editing steps. Your original footage is never touched, since Vex always works on a separate copy of the file. It can trim clips, merge them, change playback speed, add fades and text overlays, remove silent gaps, and pull out highlight segments. It also has an automatic color grading feature that samples frames from your footage and applies a matching FFmpeg filter to fix exposure, contrast, saturation, and color balance, with looks like natural, cinematic, warm, or documentary. On the audio and captions side, Vex can transcribe a video locally using Whisper, generate a transcript and subtitle file, burn those subtitles into the video, and automatically summarize long clips into shorter highlight cuts. It can turn a single video into several vertical short clips with captions and suggested B-roll footage pulled from Pexels. A newer feature called auto visuals looks at the transcript, decides which moments deserve a custom generated graphic, and renders those using either an HTML based animation engine or Manim for math heavy visuals. Projects are saved so you can resume editing later with your timeline history intact, and you can undo or redo changes. Finished videos can be exported with presets sized for YouTube, Instagram, TikTok, X, or as podcast audio, with a validation pass to confirm the export actually works. This is aimed at people who want the speed of command line tools but still want to describe edits conversationally rather than memorizing an editor's interface. It requires Python 3.11 or newer and FFmpeg installed on your system, with Node.js needed only if you want the more advanced generated visuals.
A terminal tool that edits video from plain English commands, using FFmpeg and an AI model to trim, color grade, caption, and export clips.
Mainly Python. The stack also includes Python, FFmpeg, MoviePy.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.