content-scout

stillforming

YouTube channel monitoring and daily content briefing pipeline. Monitors configured channels for new uploads, downloads videos, extracts/classifies visual frames (charts, slides, screens vs talking heads), transcribes audio, and generates a daily markdown brief with key takeaways. Use when: (1) processing YouTube videos for visual and transcript analysis, (2) generating daily content briefs from monitored channels, (3) running the content-scout pipeline or any of its steps, (4) managing channel watchlists, (5) frame extraction or classification tasks. Requires: yt-dlp, ffmpeg, Python 3.10+, PIL/Pillow, imagehash, python-slugify. Optional: OpenAI API (transcription fallback), notion-client (Notion sync).

Transcript Studio

stillforming

Deep YouTube video processing into rich Notion pages with speaker-diarized transcripts, embedded visual frames, and AI-generated summaries. Use when: (1) user asks to process, transcribe, or analyze a YouTube video, (2) creating a Notion page from a video, (3) running the transcript studio pipeline, (4) generating summaries, chapters, or shorts candidates from video content, (5) setting up a Transcript Studio Notion database. Depends on content-scout skill for frame extraction and classification steps. Requires: Apple Silicon Mac (mlx-whisper), ffmpeg, yt-dlp, Python 3.10+.