REDDIT · REDDIT// 18d agoOPENSOURCE RELEASE

NoobScribe ships local transcription, diarization, speaker memory

NoobScribe is a self-hosted, Whisper-compatible FastAPI transcription stack built on parakeet-diarized and packaged for Docker, with a web UI for recording or uploading audio and managing recordings, transcripts, and speakers. It uses Canary 1B v2 for speech-to-text, Pyannote for diarization, and Chroma-backed speaker memory so known voices keep their names across future and retroactive transcripts.

// ANALYSIS

The real product is not speech-to-text; it is transcript memory. By persisting speaker embeddings, NoobScribe makes diarization compound across meetings instead of resetting every time.

–Whisper compatibility lowers adoption friction because existing tools can point at `/v1/audio/transcriptions` without rewriting clients.
–Chroma-backed embeddings are the moat: new speakers can be matched later and old transcripts can be rewritten with better names.
–The setup tradeoff is real, with a Hugging Face token, Docker, and possible GPU dependence, but that buys local control and privacy.
–This shines most in meetings, interviews, and recurring calls where speaker attribution matters almost as much as raw transcript accuracy.

// TAGS

noobscribespeechembeddingvector-dbapidevtoolself-hostedopen-source

DISCOVERED

18d ago

2026-03-24

PUBLISHED

18d ago

2026-03-24

RELEVANCE

8/ 10

AUTHOR

meganoob1337