TTS/STT: Speech-to-Text using Gemini in Unified API

[Prajna1999]

Is your feature request related to a problem? Please describe.
Currently the unified API only supports text completions using OpenAI as the provider.
Since TTS/STT use cases are on priority, extending the llm/call endpoint to support Gemini (and others) as well audio as an extra modality as well.

Describe the solution you'd like
Start adding Gemini-2.5-Pro as a STT provider to the unified API. Integrate others simultaneosuly.

Additional context
Gemini Docs

Solution Doc