Add Google Gemini TTS API support

[shoryasethia]

Description

This PR adds support for Google Gemini's Text-to-Speech (TTS) API to Verbi, allowing users to use Google's latest TTS capabilities alongside existing providers.

Changes Made

• Added Google Gemini TTS integration to voice_assistant/text_to_speech.py
• Updated configuration to support gemini as a TTS model option
• Added API key management for Google API
• Updated requirements to include google-generativeai
• Updated documentation in README

Usage

Set TTS_MODEL = 'gemini' in config.py and add GOOGLE_API_KEY to your .env file.

References

• Google Gemini Speech Generation Documentation: https://ai.google.dev/gemini-api/docs/speech-generation

[shoryasethia]

@PromtEngineer Hello, Requesting your review for this PR.

[shoryasethia]

Hi @PromtEngineer

PR Update: Gemini TTS Support - Ready for Review

I've updated this PR with significant improvements and fixes. The implementation is now fully tested and working.

What This PR Adds

Google Gemini Text-to-Speech Integration - Adds support for Google's latest TTS capabilities using the Gemini 2.5 Flash Preview TTS model.

Major Changes & Fixes (Jan 24, 2026)

1. Migrated to New Official Google API

• Replaced deprecated google-generativeai (v0.8.6) with google-genai (v1.60.0)
• Updated all API calls to use the new genai.Client() and types classes
• Eliminates deprecation warnings

2. Fixed Deepgram SDK Compatibility

• Upgraded from v2.12.0 → v5.3.1
• Migrated transcription and TTS to use new v5 API (client.listen.v1.media.transcribe_file(), client.speak.v1.audio.generate())
• Removed deprecated imports (PrerecordedOptions, FileSource, SpeakOptions)

3. Fixed Audio Format Issues

• Added proper WAV file headers for Gemini TTS PCM audio output
• Configured for 24kHz, 16-bit, mono format
• Resolves pygame playback errors

Configurable Options

Users can customize Gemini TTS in config.py