wesbos/syntax-supercut-studio — explained in plain English
Analysis updated 2026-05-18
Search a library of podcast or YouTube transcripts for a phrase and stitch every match into one video.
Generate word-level transcripts for downloaded videos that don't have one yet.
Build a fake sentence out of individual word clips a speaker never actually said.
| wesbos/syntax-supercut-studio | omnitarium/scoptix | allstarswc/allstars | |
|---|---|---|---|
| Stars | 61 | 61 | 60 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | hard |
| Complexity | 3/5 | 3/5 | 4/5 |
| Audience | developer | developer | general |
Figures from each repo's GitHub metadata at analysis time.
Needs Node 22+, ffmpeg, aubiopitch for songify, and an XAI_API_KEY for transcription and AI regex suggestions.
Syntax Supercut Studio is a local web app for creating video supercuts from podcast or YouTube transcripts. You run it on your own machine, and it connects to folders of downloaded videos and their word-level transcripts so you can search, filter, and stitch together clips into a new video file. The core feature is the supercut builder: you type a phrase or a regular expression, and the app finds every moment in your video library where that phrase was spoken, then stitches those clips together into a single MP4. There is also an AI-assisted mode that can suggest better search patterns using an xAI API key. Beyond phrase matching, the app has several other creative tools. The sentence builder picks one clip per word from your library to assemble a sentence the speaker never actually said. The songify tool detects the pitch of each clip and arranges them into a melody. The transcribe section uses xAI speech-to-text to generate word-level transcripts for any video files that are missing one. The app organizes media into "buckets," which are folders of videos you download yourself. Each bucket has subfolders for source videos, transcripts, and rendered supercuts. Once set up, you open the app in a browser at localhost:5180, browse your library, and run any of the tools through the web interface. Finished renders are saved to disk and can be played back, downloaded, or deleted from the clips page. This is built with SvelteKit and TypeScript. It requires Node 22 or newer, plus command-line tools like ffmpeg for video processing and aubiopitch for the pitch detection used by the songify feature. The app runs entirely on your own computer, and your media files never leave your machine.
A local web app that turns your own video and transcript files into supercuts by searching for spoken phrases.
Mainly TypeScript. The stack also includes TypeScript, SvelteKit, Node.js.
The README does not state a license.
Setup difficulty is rated moderate, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.