Now in public beta!
Every voice model.
Ultra-Fast.
A fast, stable, and scalable API for top new text-to-speech models, including Sesame CSM-1B and Dia.
Vogent Voicelab is officially in public beta!
Get started in seconds
Sign up to get $5 in credits to run top voice models.
Use the latest voice models in seconds
Run the latest super-realistic voice models – like Sesame CSM-1B, Dia, Chatterbox, Orpheus, and more – with one API. Our hosted models run on an optimized voice inference stack and are post-trained to improve quality, so you can run state-of-the-art research in production. Get started in a few lines of code without managing compute.
nari-labs/dia
canopyai/orpheus
hexgrad/kokoro
resemble-ai/chatterbox
sesame/csm-1b
Better, faster, cheaper
Higher-quality than popular closed-source text-to-speech for a fraction of the price. Optimized compute for real-time inference and sub-200ms time-to-first-token, so you can use ultra-realistic models for voice agents.
Voice clone, fine-tune, and make it your own
Use zero-shot voice cloning natively, or run our fine-tune recipes to more deeply adjust style while training and hosting on our infrastructure.
Everything you need
Scale up instantly
From a single voiceover to thousands of concurrent voice agents, our infrastructure scales compute with your usage and deploys across the globe.
Secure and Compliant
SOC 2 Type II and HIPAA compliant
Enterprise-level Deployment
Use our API's, or host our inference stack on-prem or in your VPC.
Committed-use Discounts
High-volume user? Get discounts by committing to monthly usage.