epic: Nitro supports speech/hear capabilities

[hiro-v]

I would like to run TTS and STT supported in Nitro for my personal usage (self hosted Otter.ai)
For now nitro supports "think" - running LLM
This new feature is "hear"

https://platform.openai.com/docs/api-reference/audio/createTranscription

via whisper or something

[tikikun]

need to resolve the difference in ggml model + file between whisper and llama cpp

[freelerobot]

Impact: provides voice transcription