← ClaudeAtlas

interview-transcriptlisted

Download and transcribe interviews/talks from YouTube videos and X.com (Twitter) videos/Spaces. Clean transcripts with Claude (removing filler words and transcription errors while preserving facts), and send to Kindle with cover image and clickable table of contents. Use when user shares a YouTube or X.com URL and wants a transcript, or asks to transcribe a video interview.
smorrel1/claude-code-skills · ★ 2 · AI & Automation · score 65
Install: claude install-skill smorrel1/claude-code-skills
# Interview Transcript Skill Download transcripts from YouTube or X.com, clean them with AI, and send to Kindle with cover image and clickable TOC. ## Prerequisites ### Core (required for all transcripts) - **yt-dlp**: `brew install yt-dlp` or `pip install yt-dlp` (subtitle/audio extraction) - **jinja2**: `pip install jinja2` (EPUB template rendering) - **Calibre**: `brew install calibre` (provides `ebook-convert` for EPUB creation) - **Pillow**: `pip install pillow` (cover image composition and the title cover) - **pdf-to-kindle skill**: supplies `kindle_cover.py`, the shared cover drawer ### X.com fallback (only needed when no YouTube equivalent found) - **ffmpeg**: `brew install ffmpeg` (resample audio to 16kHz mono for whisper) - **whisper-cpp**: `brew install whisper-cpp` (audio transcription, binary at `/usr/local/opt/whisper-cpp/bin/whisper-cli`) - **Whisper model**: e.g. `ggml-small.bin` from superwhisper app or [huggingface.co/ggerganov/whisper.cpp](https://huggingface.co/ggerganov/whisper.cpp/tree/main) ### Optional services - **Gmail `email` skill** with OAuth credentials for Kindle delivery - **Google Drive** OAuth token (`~/.claude/skills/email/token_drive.json`) for shareable link generation ## Supported Sources | Source | URL Pattern | Method | |--------|------------|--------| | YouTube | `youtube.com/watch?v=...`, `youtu.be/...` | Subtitle extraction (SRT) | | X.com / Twitter | `x.com/.../status/...`, `twitter.com/...` | Subtitles if available, otherwis