January 29, 2026
Bees vs Buffering
Show HN: Transcribee: YouTube transcriber that builds a knowledge base
HN hive buzzes over a tool that turns videos into your personal brain
TLDR: An open-source Mac app called Transcribee auto-transcribes and organizes videos into a searchable library. The crowd loves the “instant notes” promise but debates privacy, API costs, and Mac-only limits—fans call it productivity gold, skeptics want cross-platform and fewer third-party dependencies.
Meet Transcribee, the open‑source Mac app that slurps up YouTube, TikTok, Reels, and podcasts, then spits out neat, speaker‑labeled notes that self‑organize into folders. The community reaction? Split right down the honeycomb. Fans cheered the “watch less, learn more” vibe, loving how it auto‑sorts your transcripts and makes them easy to paste into LLMs (large language models) like ChatGPT and Claude. Skeptics buzzed about privacy and cost—it needs API keys for ElevenLabs (for transcription) and Anthropic (for categorizing), which had some folks side‑eyeing whether their binge‑watch habits should run through third‑party services.
The biggest drama: macOS‑only. Windows and Linux users showed up with pitchforks made of bash scripts, begging for cross‑platform support while Mac users yelled “Works on my machine!” Integration with Clawdbot (an AI assistant) sparked mini‑wars: some called it agent overkill, others said it’s perfect for hands‑free “transcribe this” life. There were jokes galore—bee puns everywhere, plus memes about turning your chaotic watch history into a librarian‑approved archive. One commenter bragged they finally have a folder for “Ilya Sutskever on AGI,” while another joked they can now summarize podcasts they never finished. The vibe: productivity nerds in love, privacy hawks on alert, and platform purists sharpening their stingers.
Key Points
- •Transcribee is an open-source macOS tool to transcribe YouTube, Instagram Reels, TikTok, and local media.
- •It uses ElevenLabs (scribe_v1_experimental) for transcription with speaker diarization and Claude for auto-categorization.
- •Outputs include a speaker-labeled transcript.txt, metadata.json, and optional raw JSON with word-level timestamps.
- •Installation integrates with Clawdbot via ClawdHub or manual GitHub cloning; dependencies include yt-dlp, ffmpeg, pnpm, Node.js 18+.
- •Supports common audio/video formats and URLs; tested on macOS Sonoma and designed to build a self-organizing knowledge library.