​

Voice cloning update

We’re streamlining our voice cloning experience! After continued model improvements, we’ve decided to focus exclusively on our instant voice cloning technology! Why? Because it’s faster, more accurate, and just plain better. The professional voice cloning option has been retired, but don’t worry - you won’t miss it. Our instant clones are now so good, they’ll make you forget the old way ever existed.

• Existing professional voice clones have been converted to instant voice clones. All of your existing voices will continue to work.
• Instant clones achieve better speaker similarity while requiring significantly less source audio.
• As always, users can create unlimited voice clones.

For more details on our voice cloning capabilities, check out our Voice Cloning guide.

​

Goodbye, Aurora. Hello, Blizzard!

Today marks an important milestone: Blizzard is now ready for production use! To provide the best possible experience for all users, we’ve made the following changes:

• Blizzard is now our default model
• The Aurora model ID (aurora) automatically routes to Blizzard

This consolidation ensures consistent high performance and reliability across all speech synthesis requests. For more details, check out our Models guide.

​

Experimental model release: Blizzard

We are excited to announce the release of our experimental model, Blizzard! Try it out in the Playground and in the API.

Learn more about how to use this model here.