docs(custom-stt): fix the documentation about how to send segments with custom stt

[42atomys]

This pull request updates the documentation for Omi's real-time audio transcription system, to add a comprehensive section explaining how to integrate your own WebSocket-based transcription/diarization service, including message formats, requirements, and example data flows.

[gemini-code-assist]

Code Review

This pull request adds valuable documentation for integrating a custom STT service. The new section is well-structured and provides a good overview. My review includes two suggestions to improve clarity and accuracy. One comment addresses an ambiguous requirement for the custom service's response format, and the other points out a field in an example payload that is ignored by the backend. Fixing these will make the documentation more precise for developers.

[mdmohsin7]

Don't think this is the right place for custom stt because it never touches the backend and this file is primarily related to how the omi's stt system works

[42atomys]

Hi @mdmohsin7 as no real place exist for custom STT, maybe create a new section, if you have any directive of where to locate it, I will create the section and move this doc on it

Waiting your feedback to change

[beastoin]

Ah, no worries. His updates let the dev know they can send the suggested transcript segment to the backend. Good first steps, then we can clarify that the dev can create/share their configs via the Omi app.

[beastoin]

lgtm @42atomys