A curated registry of tools, models, microservices, and integrations for text-to-speech (TTS) with NVIDIA technology. Covers cloud APIs, self-hosted NVIDIA Speech NIM Microservices, embedded Jetson deployments, voice cloning, and agent-facing MCP integrations.
- Core NVIDIA TTS Products
- TTS NIM Microservice
- NVIDIA Riva TTS
- Riva Studio / NeMo TTS
- Digital-Human Audio (ACE / Audio2Face)
- Deployment Targets
- Voice Capabilities
- SDKs, Clients & Integrations
- Community Tools & Examples
- Learning Resources
- Contributing
- License
| Product | Best For | Deployment |
|---|---|---|
| NVIDIA Speech NIM Microservices | Production self-hosted TTS microservice | Cloud / data center / workstation GPUs |
| NVIDIA Riva TTS | Full speech-AI stack (ASR + TTS + NMT) | Self-hosted Docker or Jetson |
| Riva Studio / NeMo TTS | Custom voice training and fine-tuning | Cloud training, on-prem GPUs |
| NVIDIA ACE / Audio2Face | Digital-human facial animation driven by TTS audio | RTX workstations, cloud |
The NVIDIA Speech NIM Microservices package pre-trained NeMo models into optimized, self-contained inference containers.
| Model | Languages | Modes | Key Capability |
|---|---|---|---|
| Magpie TTS Multilingual | English, Spanish, French, German, Mandarin, Vietnamese, Italian, Hindi, Japanese | Streaming + Offline | Multi-voice, multi-language synthesis |
| Magpie TTS Zeroshot | English | Streaming + Offline | Voice cloning from 3–10s reference audio |
| Magpie TTS Flow | English | Offline | Voice cloning with audio prompt + transcript |
| Chatterbox TTS Multilingual | 23 languages incl. Arabic, Korean, Portuguese, Russian | Streaming + Offline | Emotion exaggeration |
- TTS NIM Documentation
Official - TTS NIM Tutorial
Official - Speech NIM models
Official— per-model NIM catalog and containers - Chatterbox TTS Multilingual
Community - Chatterbox TTS NIM deployment docs
Official - Quick run (requires NGC login / NVAIE entitlement):
docker run --gpus all --runtime=nvidia -it \ -e NVIDIA_API_KEY=$NVIDIA_API_KEY \ -p 50051:50051 \ nvcr.io/nim/nvidia/tts:latest
Riva is NVIDIA's speech-AI SDK. The TTS skill supports high-quality, low-latency synthesis with gRPC and Python clients.
- NVIDIA Riva
Official— Speech AI SDK with ASR, TTS, and NMT. - Riva TTS Quick Start
Official - Riva Skills on Embedded
Official— Jetson deployment guide. - Riva Python Client
Official - List voices & synthesize:
python3 /opt/riva/examples/talk.py --list-voices python3 /opt/riva/examples/talk.py \ --text "Hello from NVIDIA Riva" \ --voice Magpie-Multilingual.EN-US.Mia \ --language-code en-US \ --output /opt/riva/wav/output.wav
For custom voices and fine-tuned models.
- NVIDIA Riva Studio
Official— No-code voice customization, recording, training, and deployment. - NVIDIA NeMo
Official— End-to-end framework for building and customizing speech and language models. - NVIDIA NeMo Speech
Official— Dedicated repository for ASR and TTS research and pre-trained models (split from the main NeMo project). - NeMo TTS Models
Official— Tacotron2, FastPitch, RadTTS, Mixer-TTS, and more. - NeMo Curator
Official— Data-curation tools for training custom TTS datasets. - MagpieTTS multilingual open-weight checkpoint
Official/Open— Open-weight MagpieTTS multilingual 357M checkpoint. - MagpieTTS multilingual demo
Official— Hugging Face demo space for MagpieTTS multilingual.
While not pure TTS, these products consume TTS output to animate digital humans.
- NVIDIA ACE
Official— Avatar Cloud Engine combining Riva ASR/TTS, NeMo LLMs, and Audio2Face animation. - NVIDIA Audio2Face
Official— Generate real-time 3D facial animation from audio. - Audio2Face NIM
Official— Self-hosted microservice for facial animation on RTX. - NVIDIA Nemotron 3 VoiceChat
Official— Speech-to-speech conversational voice AI; early access via developer.nvidia.com. Blog: Building NVIDIA Nemotron 3 Agents.
End-to-end voice agents and blueprints built on NVIDIA speech models.
- NVIDIA voice-agent-examples
Official— Reference voice-agent pipelines using NemotronASRService / NemotronTTSService (renamed from RivaASRService / RivaTTSService as of v0.4.0). - NVIDIA Nemotron Voice Agent Blueprint
Official— Build deployable voice agents with NVIDIA Nemotron speech and reasoning models. Repository: NVIDIA-AI-Blueprints/nemotron-voice-agent. Container: catalog.ngc.nvidia.com/nvidia/blueprint/nemotron-voice-agent.
| Target | Product | Notes |
|---|---|---|
| Cloud / data center | Speech NIM, Riva Server | Requires NVAIE for production; Docker/Helm on Kubernetes. Note: Riva Server x86 data-center deployments are deprecated as of Riva 2.24.0; use NVIDIA Speech NIM for x86/cloud TTS. |
| Workstation / RTX PC | Speech NIM, Riva, Audio2Face | Local inference for dev/test and digital humans |
| Embedded / Jetson | Riva TTS (arm64 quickstart) | Optimized models for Jetson Orin / Nano |
| Hosted API | build.nvidia.com NIM endpoints | Prototyping and low-volume inference; rate-limited |
- Multilingual synthesis — 23+ languages across Magpie and Chatterbox models.
- Voice cloning / Zeroshot — Clone a voice from a few seconds of reference audio.
- Emotional styles — Select emotions such as Happy, Calm, Excited via voice naming (e.g.
Magpie-Multilingual.EN-US.Aria.Happy). - SSML support — Control phoneme pronunciation and prosody.
- Streaming & offline — Low-latency streaming for agents; offline for batch content.
- Custom pronunciation dictionaries — IPA-based word-to-phoneme mappings.
- Riva Python Client
Official— Official Python gRPC client for Riva TTS. - Riva C++ / gRPC API
Official— Low-level gRPC definitions. - LiveKit NVIDIA Riva TTS plugin
Community— Use Riva TTS in LiveKit voice agents.- Install:
uv add "livekit-agents[nvidia]~=1.5"
- Install:
- NVIDIA Pipecat voice-agent-examples
Official— Voice-agent examples using Pipecat with NemotronASRService / NemotronTTSService (renamed from RivaASRService / RivaTTSService as of v0.4.0). - OpenAI-compatible NIM endpoints
Official— TTS NIM exposes standard HTTP/gRPC APIs.
- **LLMAvatarTalk-An-Interactive-AI-Assistant
Community— Tutorial project combining Riva ASR/TTS, NIM LLM, Audio2Face, and Unreal Engine MetaHuman. - livekit-plugins-nvidia
Community— LiveKit plugin for NVIDIA Riva TTS.
- NVIDIA Speech NIM Microservices Docs
Official - Riva TTS Docs
Official - NeMo TTS Docs
Official - NVIDIA ACE / Digital Humans
Official - NVIDIA Speech AI Forum
Official
Read CONTRIBUTING.md for the quality bar, entry format, and PR process.
This list is released into the public domain under CC0-1.0.
Enterprise AI Atlas is maintained by Vibe Coding Agency. We prototype and ship agentic systems, MCP servers, and enterprise AI integrations for teams that need working software fast — without hiring a full AI engineering team.
Free guide: The Non-Technical Founder's Guide to Agentic AI — what agents and MCP servers are, and how to get a system built.