[Open Source] Maya1 TTS for Unity (Fully Local, No External Apps, Unlimited Expressive Voice Design)

Community Showcases
, , ,
1

Hey everyone! I wanted to share something that changes the game for NPC interactions and accessibility in Unity. If you’ve been looking for ElevenLabs-quality local speech in Unity without the lag or API costs, you need to check out Maya1 TTS.

Most local AI solutions require a separate backend. This runs directly inside Unity via Llama.cpp and ONNX, making it a “drop-in” solution for NPCs.

:performing_arts: The Killer Feature: Natural Voice Design

Unlike other TTS where you pick from a list of voices, Maya1 lets you “prompt” a voice. Instead of selecting “Voice_01,” you just describe it:

Demon character, Male voice in their 30s with a Middle Eastern accent. screaming tone at high intensity.

The model understands the natural language description and generates a unique voice profile on the fly—no cloning or training data needed.

:rocket: Highlights:

  • Expressive Emotion Tags: Trigger <laugh>, <whisper>, <rage>, or <gasp> mid-sentence for human-level performance.
  • Ultra-Low Latency: Sub-100ms response. It supports token-to-audio streaming, meaning your NPCs can start talking before the full sentence is even finished.
  • 100% Private & Local: No API keys, no subscriptions, and zero data leaves the device.
  • Single GPU Ready: Runs comfortably on consumer cards (6GB VRAM for quantized models).
  • Commercial Friendly: Released under Apache 2.0/Commercial-ready terms—no royalties or “non-commercial” handcuffs for the core engine.

Repo: maya-tts-unity

:mobile_phone: Need a Mobile Solution?

If you’re targeting mobile hardware instead of desktop, check out my other specialized Unity integrations:

  • NeuTTS for Unity: The go-to for high-end mobile devices. It supports instant voice cloning and high-fidelity output optimized for modern ARM processors.Repo: neutts-unity
  • Vyvo TTS for Unity: Optimized for mid-range devices. This is an experimental framework designed to bring neural LLM-based speech to more modest mobile hardware.Repo: vyvo-tts-unity

Give it a spin and let me know how it handles your character descriptions! I’m looking for feedback from the community, so if you encounter bugs or have suggestions, please open an issue on the respective GitHub pages!

1 Like
2

If you’ve seen a newer or better TTS model come out and you want me to make it available in Unity, just comment below!

Requirement: I can integrate almost any model as long as the base model is available in GGUF format and the decoder is available in ONNX format.

1 Like