Scribe

PrivateStage 2-11

Long videos to .txt and .srt transcripts on your own GPU, with no API keys.

Fig 1 / 1The job queue, and an SRT it produced
Upload chunk
4 MB
Service
~125 lines

Private repository.

  • Python
  • FastAPI
  • faster-whisper
  • CUDA

About

A small FastAPI service with one upload page: a background worker transcribes recordings with faster-whisper and offers plain-text and subtitle downloads.

Highlights

  • Streams multi-GB uploads to disk in 4 MB chunks, hashing in the same pass, so a file is never held in memory.
  • Transcripts are cached by content hash and model, so a repeated upload returns instantly.
  • CUDA float16 first, CPU int8 as the fallback, with NVIDIA libraries preloaded from pip wheels.

How it flows

From start to goal, the path the data takes

  1. Upload · 4 MB chunks
  2. Hash cache
  3. faster-whisper
  4. .txt + .srt