NVIDIA Opens Magpie TTS for Multilingual Voice Agents
The open-weights text-to-speech model targets low-latency speech synthesis with full control over deployment.
NVIDIA has released Magpie TTS, an open-weights text-to-speech model aimed at developers building multilingual voice agents. According to the company's announcement on Hugging Face, the model is designed for low-latency speech synthesis with full control over how and where it runs.
The pitch is straightforward: voice agents that respond in real time need synthesis that is both fast and deployable on the developer's own terms. By publishing the weights, NVIDIA lets teams host the model in their own environments rather than routing every request through a hosted API — a meaningful difference for latency-sensitive and privacy-conscious applications.
Why it matters
Text-to-speech is one of the last links in the conversational AI chain, and it is often where hosted services introduce delay, cost, and data-handling constraints. An open-weights, multilingual option gives builders more room to maneuver:
- Low-latency synthesis suited to interactive voice agents
- Multilingual coverage for broader deployments
- Self-hosting for tighter control over infrastructure and data
NVIDIA is releasing Magpie TTS under a custom license rather than a standard permissive one, so teams will want to review the terms before building on it. As an initial release, it establishes a foundation that NVIDIA can iterate on — and adds another serious open contender to a TTS landscape that has moved quickly toward openly available models.
Sources
More from NVIDIA
All NVIDIA releases →NVIDIA's Cosmos 3 Brings Open Interactive World Models
The new open release pushes beyond passive video generation toward simulated worlds that respond to input.
NVIDIA Releases Cosmos 3 for World Simulation
The new open models push NVIDIA's world-foundation effort toward interactive, real-time video generation.

NVIDIA releases compact Dory 1.0 4B text model
The new Nemotron Labs model packs a 4-billion-parameter language model trained on NVIDIA's Nemotron datasets.
More in Text → Speech
All Text → Speech →Haru-4B Brings Full-Duplex Speech to Six Languages
A compact 4-billion-parameter model handles real-time, two-way voice conversation across six languages in a single system.

Zyphra releases Zonos3-Mini multilingual TTS
The compact open text-to-speech model supports multiple languages and voice cloning from short audio samples.
OpenBMB's Phonon Brings High-Quality TTS On-Device
The new text-to-speech model promises benchmark-topping voice synthesis that runs locally rather than in the cloud.
0 comments
No comments yet. Be the first to weigh in.