[Docs]: STT custom provider Query parameters, request rules, and response rule

[rtATvw]

Documentation area

docs/integrations/stt/custom-stt

What is missing or unclear?

The custom provider isn't documented for example what needs to be set to send the json object
{"mode": "online", "wav_name": "wav_name", "is_speaking": true, "wav_format":"pcm", "audio_fs":{"$var":"sample_rate"},"chunk_size":[5,10,5],"itn":true}
At the beginning
The default request rules appear to send raw PCM frames which is fine.
However sending {"is_speaking": false} at the end of audio isn't clear.
And the response rules for parsing the text variable out of the object {"mode": "2pass-online", "wav_name": "wav_name", "text": "asr ouputs", "is_final": True, "timestamp":"[[100,200], [200,500]]", "stamp_sents":[]}
aren't clear

Trying to extract it from:

voice-ai/ui/src/providers/custom-stt/contract.ts

Line 22 in 2a946d8

export const CUSTOM_STT_DSL_VARIABLES = [

voice-ai/ui/src/app/components/providers/websocket-dsl-editor/suggestions.ts

Line 12 in 2a946d8

CUSTOM_STT_DSL_VARIABLES,

and

voice-ai/ui/src/providers/custom-stt/__tests__/template.test.ts

Line 3 in 2a946d8

CUSTOM_STT_DSL_VARIABLES,

is less than satisfying.

What would you expect instead?

Some examples of how to configure the custom provider

References or examples (optional)

https://github.com/modelscope/FunASR/blob/21e6ededcc5dd83cce5044ccff44091f75ee524e/runtime/docs/websocket_protocol.md?plain=1

Contribution

• I would like to work on this issue
• I can submit a PR with documentation changes

[iamprashant]

Thanks @rtATvw Will be updating documentation and supported DSL for stt/tts.

[iamprashant]

We have updated our documentation for custom STT/TTS. Please let us know if any further improvements are needed.

Speech to text : https://doc.rapida.ai/integrations/stt/custom
Text to speech : https://doc.rapida.ai/integrations/tts/custom

[rtATvw]

The "when": { "packet": "turn_change" } request rule isn't triggered on session-start.