Table of Contents
What it does
VidPipe is a CLI tool that takes raw video recordings and runs them through a 14-stage automated pipeline: ingestion, transcription, silence removal, captioning, shorts extraction, medium clips, chapter detection, summaries, social posts, and queue building. The output is ready-to-publish content for multiple platforms.
Language
TypeScript (97.9% of the codebase).
Install
npm install -g vidpipe
Prerequisites: Node.js 20+, FFmpeg 6.0+, OpenAI API key.
Key features
- AI transcription using OpenAI Whisper with word-level timestamps
- Silence removal - context-aware, capped at 20%
- Karaoke-style captions with word highlighting
- Short-form clips (15-60s) in multiple formats for TikTok, Reels, Shorts
- Medium clips (1-3 min) with transitions
- Chapter detection in multiple formats
- Social media posts tailored for TikTok, YouTube, Instagram, LinkedIn, X
- Blog post generation in Dev.to style
- Face detection for webcam cropping
- Content ideation backed by trend research
CLI commands
| Command | Description |
|---|---|
vidpipe [video-path] |
Process a single video |
vidpipe --watch-dir |
Monitor a folder for new recordings |
vidpipe ideate |
Generate trend-backed content ideas |
vidpipe review |
Launch web-based post review interface |
vidpipe doctor |
Verify prerequisites |
Value
Good for content creators who record long-form and need automated repurposing. The watch directory mode is nice for a "record and forget" workflow. The 14-stage pipeline covers basically everything you'd need to go from raw footage to published content across platforms.