How do I use enhanced tier of any model in a live transcription (websocket connection)?

[yashpalzala]

I'm trying to use enhanced tier of general model but I get error saying :

[ERROR:flutter/runtime/dart_vm_initializer.cc(41)] Unhandled Exception: WebSocketChannelException: WebSocketChannelException: WebSocketException: Connection to 'https://api.deepgram.com:0/v1/listen?encoding=linear16&sample_rate=16000&language=en-IN&speech_final=true&model=general-enhanced#' was not upgraded to websocket

My server url along with parameters is "wss://api.deepgram.com/v1/listen?encoding=linear16&sample_rate=16000&language=en-IN&speech_final=true&model=general&tier=enhanced"

The problem only occurs if I add the "tier" parameter.

[briancbarrow]

Hi @yashpalzala the speech_final parameter is something that will show up in the response object, not something that should be passed into the request.

Do you still get errors if that is removed?

I just tested it using

[yashpalzala]

speech_final parameter is not the one causing the problem I removed it or keep it does not affect the outcome but when I put tier=enhanced as a parameter it causes the above given error.

And i don't get error if I remove the tier parameter, but then how do I use enhanced tier of any model.

I have followed the documentation here :

By default, Deepgram applies its Base tier. If you would like to use a different tier, for a hosted deployment, add a tier parameter in the query string as well:

tier=enhanced&model=OPTION

Link to the documentation page: https://developers.deepgram.com/documentation/features/model/

[briancbarrow]

Hi @yashpalzala, Sorry I didn't catch this before now, but I just noticed you are using the en-IN language. Currently only the base tier is supported with that language code.