llsourcell/stuff — explained in plain English
Analysis updated 2026-07-29 · repo last pushed 2026-01-12
Run a faceless YouTube Shorts channel about science topics on autopilot.
Generate a backlog of short science videos in batch without manual editing.
Test the full video creation pipeline with a dry run before publishing.
| llsourcell/stuff | 0-bingwu-0/live-interpreter | 010zx00x1/faresnipe | |
|---|---|---|---|
| Stars | 2 | 2 | 2 |
| Language | Python | Python | Python |
| Last pushed | 2026-01-12 | — | — |
| Maintenance | Quiet | — | — |
| Setup difficulty | hard | moderate | easy |
| Complexity | 4/5 | 2/5 | 2/5 |
| Audience | pm founder | general | general |
Figures from each repo's GitHub metadata at analysis time.
Requires paid API keys for OpenAI, ElevenLabs, and Replicate, plus an Ubuntu server configured for scheduled task execution.
Inside You is an automated content factory for YouTube Shorts. It takes a science topic, writes a script, generates a voiceover, creates matching visuals, stitches everything together with subtitles and sound effects, and then uploads the finished video straight to YouTube. The whole pipeline is designed to run on autopilot, so you can set it up on a server and have it produce and publish short-form science videos on a regular schedule without any manual editing. The pipeline chains together several AI services. OpenAI's GPT-4 writes the script in a structured format so the system knows exactly which visuals and narration go together. ElevenLabs generates the voiceover from that script, and Replicate creates images using AI art models. FFmpeg, a standard video tool, then assembles all of these pieces into a vertical video with a heads-up display overlay, a scanning sound effect, and subtitles burned into the frame. Before anything gets uploaded, an automated quality check reviews the video for loudness, transcript accuracy, visual relevance, and safety, and you can adjust what score counts as a pass. Everything is logged in a local database so you can track what's been made. The target user is someone who wants to build a YouTube Shorts channel on science topics without spending hours writing, recording, editing, and uploading each video. You could use it to run a faceless channel about things like dust mites, space anomalies, or body biology, publishing once or twice a day. It is built for people comfortable running a Python project on a cloud server, since it is designed to live on an Ubuntu machine and run through scheduled tasks. What stands out is how hands-off it becomes once configured. You pick the topics from a file, set the schedule, and the system handles the rest. You can test the full pipeline without uploading by running a dry run, or generate a backlog of several videos at once. The tradeoff is that it depends on multiple paid AI services, so each video costs money in API calls. You are also limited to the specific visual style baked into the template, though you can swap out the overlay graphic and sound effects if you want a different look.
An automated pipeline that turns a science topic into a finished YouTube Shorts video with AI-written scripts, voiceovers, AI-generated visuals, subtitles, and automatic uploads on a schedule.
Mainly Python. The stack also includes Python, OpenAI GPT-4, ElevenLabs.
Quiet — no commits in 6-12 months (last push 2026-01-12).
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly pm founder.
This repo across BitVibe Labs
double-check against the repo, no cap.