May 24

1. 23:53

   fix: remove forced language param - causes hallucination on tiny model

   Yash030
   a6c3607
2. 23:43

   fix: raise no_speech_threshold to 0.95 for lenient segment filtering

   Yash030
   d25f1fc
3. 23:40

   fix: add segment-level logging, relax no_speech_threshold

   Yash030
   b50f9af
4. 23:35

   fix: normalize audio to [-1,1] range before Whisper

   Yash030
   979c7b2
5. 23:30

   fix: resample audio to 16kHz, add debug logging

   Yash030
   0f0d22a
6. 23:15

   fix: add python-multipart dep for file upload

   Yash030
   3399a47
7. 23:10

   Initial deploy: faster-whisper FastAPI server⁠

   Yash030 +1: Claude Opus 4.7
   0e2e8ef
8. 23:07

   initial commit

   Yash030
   Verified signature

   9f6dbee