This message was deleted.
# helpdesks
steep-balloon-41261
08/14/2023, 2:26 PMThis message was deleted.
m
magnificent-art-43333
08/14/2023, 2:30 PMAre you trying to do something similar to this?
https://github.com/livekit-examples/kitt
m
microscopic-mechanic-41538
08/14/2023, 2:54 PMYes, only instead of chatgpt, a regular user and working with their audio stream.
m
magnificent-art-43333
08/14/2023, 3:10 PMSure thing @microscopic-mechanic-41538 — you should be able to repurpose that code to accomplish what you need. The LK GPT service is where the main parts you’re interested in reside: https://github.com/livekit-examples/kitt/tree/main/lkgpt-service
magnificent-art-43333
08/14/2023, 3:13 PMThere’s also a new example up for Whisper STT, but we still need to add a facility to the Python SDK making it easy to publish synthesized speech back into the room. From what I understand of your use case, using the Go SDK is your best bet for the next couple weeks.
🔥 1
m
microscopic-mechanic-41538
08/18/2023, 8:58 AMAs I understand it I need to write speech to speech service inside go sdk, create a new thread with synthesized speech and call it on client side.
Could you tell me where the audio stream is generated? At the stage of reading from the microphone and passing packets to the buffer.
m
magnificent-art-43333
08/18/2023, 2:52 PM@microscopic-mechanic-41538 which part are you looking for, where speech is transformed into text, or text into speech?
magnificent-art-43333
08/18/2023, 2:56 PMgptparticipant.go