Video Ingest
Transforms video and audio recordings into structured knowledge notes using scene-aware keyframes and Whisper transcripts.
Video Ingest enables Claude Code to convert local recordings or online video URLs (YouTube, TikTok, Instagram) into comprehensive vault documentation using the claude-real-video (crv) CLI. It extracts deduplicated scene keyframes alongside OpenAI Whisper transcriptions, correlates visual screens with spoken topics, redacts sensitive secrets, and compiles the content into structured raw archives and interconnected markdown wiki entries with embedded screenshots.
Key Features
01Enforces secret redaction for API keys, credentials, and sensitive screen data
02Compiles structured Raw source notes and interconnected wiki documentation
03Supports local media files and direct URLs from platforms like YouTube and TikTok
04Correlates visual scene changes chronologically with spoken transcript sections
05Extracts deduplicated visual keyframes and Whisper audio transcriptions via crv CLI
0668 GitHub stars
Use Cases
01Analyzing reference and tutorial videos from social media without manually taking screenshots
02Ingesting meeting recordings, workshops, and webinars into an Obsidian knowledge base
03Transcribing and documenting technical walkthroughs, coding sessions, or architecture demos