retrogtx/hokou-shorts-bot — explained in plain English
Analysis updated 2026-05-18
Automatically build a YouTube Shorts channel around a niche video search query.
Batch trim and reformat downloaded clips into vertical video with a branded outro.
Run a scheduled daily upload job without manually logging into YouTube each time.
| retrogtx/hokou-shorts-bot | atomicstrata/atomicmemory | azerdsq131/mcpm | |
|---|---|---|---|
| Stars | 6 | 6 | 6 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 3/5 | 4/5 | 2/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires yt-dlp and ffmpeg installed separately plus YouTube OAuth credentials.
hokou-shorts-bot is a command line automation tool that finds short YouTube videos matching a search query, trims and reformats them, appends a custom call to action outro, and uploads the results to a YouTube channel as Shorts on a recurring schedule. The pipeline has three stages. First, the scrape step uses yt-dlp, a widely used command line video downloader, to pull the top results for a configurable search query, defaulting to "satisfying 3d animation", filtering for clips between 3 and 65 seconds long. Second, the process step uses ffmpeg, a video processing tool, to normalize each clip to a vertical 1080x1920 format at 30 frames per second, trim it to a set duration (defaulting to 2 seconds), and splice on a pre made outro video that you provide. Third, the upload step picks the oldest unprocessed clips from the output folder and pushes them to YouTube via the official YouTube Data API, tracking what has already been posted in a local file called posted.json. Authentication with YouTube uses the OAuth flow: you run a one time auth command that walks you through authorization and gives you a refresh token to store in a configuration file. From that point on, uploads happen without further login prompts. For automated scheduling, the project includes a macOS LaunchAgent configuration file that runs the upload step daily at 9:00 AM. You install it by copying the file into the system's LaunchAgents folder. All raw downloads and intermediate files are kept out of version control, only the posting ledger is tracked. The project is written in TypeScript and requires yt-dlp and ffmpeg to be installed separately.
A tool that automatically finds short videos, trims and rebrands them, then uploads them to YouTube as Shorts on a schedule.
Mainly TypeScript. The stack also includes TypeScript, Node.js, yt-dlp.
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.