ROIpad ← Back to Search
news.ycombinator.com › AI insight

Insight for: Show HN: Mcptube – Karpathy's LLM Wiki idea applied to YouTube videos

Mcptube (v2/mcptube-vision), an application of Karpathy's LLM Wiki idea to YouTube videos. It extracts transcripts, detects scene changes, describes key frames with a vision model, and creates structured wiki pages for Q&A and search.
Analyzed: Apr 14, 2026
The challenge of extracting actionable intelligence from long-form video content, particularly educational or technical lectures, is a significant productivity bottleneck. Mcptube addresses this by transforming YouTube videos into structured, searchable wiki pages, leveraging vision models and transcript analysis. This approach moves beyond simple keyword search, enabling knowledge compounding and efficient Q&A. The shift from re-searching raw chunks to pre-processed, structured knowledge is a critical architectural improvement, enhancing retrieval speed and accuracy. Positioning as both a CLI/MCP server and a future SaaS platform indicates a clear commercialization path, targeting teams and organizations that rely heavily on video-based learning and knowledge sharing. This tool directly impacts corporate training, research, and content consumption efficiency.
LLM Wiki pattern YouTube videos transcript search Q&A MCP server raw chunks ingest time extracts transcripts detects scene changes ffmpeg describes key frames vision model structured wiki pages knowledge compounds FTS5 two-stage agent narrow then reason retrieval CLI (BYOK) API key SaaS platform playlist ingestion team wikis architecture tradeoffs FTS5 vs vectors file-based wiki vs DB scene-change vs fixed-interval sampling pip install mcptube