feat: Adiciona funcionalidades de Speech-to-Text e Text-to-Speech

[alguemaiYT]

Adiciona novos endpoints para STT e TTS, permitindo a conversão de áudio em texto e vice-versa.

• Cria o roteador /api/v1/stt para transcrição de áudio.
• Cria o roteador /api/v1/tts para síntese de voz.
• Estende o XAIClient para dar suporte a essas novas funcionalidades.
• Adiciona os schemas necessários para as requisições de TTS.
• Integra os novos roteadores na aplicação principal.

[bigsk1]

Thank you for the contribution, but there's a fundamental problem with this approach:

xAI does not offer STT (Speech-to-Text) or TTS (Text-to-Speech) services. This PR implements these features using OpenAI's API, which creates confusion about what services xAI actually provides.

Issues:
The PR description claims to "extend XAIClient" for STT/TTS, but these aren't xAI capabilities
Implementation uses OpenAI's Whisper and TTS APIs, not xAI
This misleads users about xAI's actual service offerings
Creates a mixed API that's not clearly documented

Recommendation:
If STT/TTS functionality is desired, it should be:
Clearly documented as OpenAI integration, not xAI extension
Implemented as separate endpoints (e.g., /api/v1/openai/stt)
Have clear documentation that these features require OpenAI API keys
Not presented as xAI functionality

Please clarify the intent and consider restructuring to avoid confusion about which services come from which provider.
Bottom line: This PR is architecturally wrong because it's trying to add non-existent xAI features using a different API provider.

[alguemaiYT]

Oh yeah, I'm using your repo as an SDK for my project, I'm planning to integrate Deepgram, Porcupine, and Grok as the LLM. Your FastAPI implementation helped me a lot — thanks!

[alguemaiYT]

Jules Bot That i tried to use add some wrong comments about my scripts, and added unwanted scripts too ;-;