Setup High-Quality Open-Source TTS in my server

I'm looking for an experienced developer to install a high-quality open-source Text-to-Speech (TTS) system on my Ubuntu server. The TTS will be used in my web app to generate audio outputs.

Requirements:
- TTS should support Brazilian Portuguese with both masculine and feminine voices.
- Must be callable via HTTP to integrate seamlessly with my app.
- Should provide audio quality comparable to 11labs.

I previously tried using Piper and Coqui xtts_v2, but neither met my quality expectations. I'm open to other models and suggestions.

Ideal Skills and Experience:
- Proficiency in setting up TTS systems on Ubuntu
- Strong knowledge of open-source TTS models
- Experience with Brazilian Portuguese language processing
- Ability to configure HTTP requests for TTS services