Send Byte audio data from call API to Whisper

[BakingBrains]

I am trying to transcribe byte data coming from twilio call stream. I followed this blog, https://www.twilio.com/blog/transcribe-phone-calls-text-real-time-twilio-vosk I want to replace vosk with whisper. I only doubt I have is regarding the input byte.

I tried with load_audio method mentioned here (#380) still no use.

audio = load_audio(audio)
customaudio_data = np.frombuffer(audio, np.int16).flatten().astype(np.float32) / 32768.0
text, _, _ = model.transcribe(customaudio_data)

Any suggestions?

Thanks and Regards

[BakingBrains]

Any suggestions here?

Thank You.

[justinlevi]

I have been working on something similar. Just got this working here;
collabora/WhisperLive#40