Google Text-to-Speech is not supported as a native voice provider in Vapi Server SDK

[Pro-Sifat-Hasan]

Summary

While integrating Vapi Server SDK for a voice assistant, I noticed that Google Text-to-Speech (TTS) is not available as a native voice.provider, even though Google is supported for transcription (`transcriber.provider = "google").
This creates confusion for developers, especially since Google already provides a fully managed TTS API with public endpoints.

Expected Behavior

Developers should be able to configure Google TTS directly, for example:

voice: {
  "provider": "google",
  "voiceId": "bn-BD-Standard-A"
}

without needing to deploy an additional middleware or custom adapter.

Actual Behavior

voice.provider = "google" is rejected with a validation error:`

  voice.provider must be one of the following values: ...

The only way to use Google TTS is via custom-voice, which requires:

• Deploying an adapter server
• Manually handling request/response transformation
• Streaming raw PCM audio

This defeats the purpose of using an existing managed TTS service with a public endpoint.

Impact

• Additional infrastructure and maintenance for developers
• Increased complexity for simple use cases
• Confusion since Google is supported for STT but not TTS
• Harder adoption for regions where Google voices (e.g., Bangla bn-BD) are higher quality than other providers

Suggestion

Consider adding native Google TTS support similar to Azure / ElevenLabs, or clearly document why Google TTS cannot be supported directly due to protocol or streaming constraints.

Even limited support (non-streaming or buffered) would significantly improve developer experience.