Company
About DuoWave
We are building a hosted API for voice AI that sounds like a person: models that listen and talk at the same time.
What we do
Most voice assistants take turns. They wait for you to finish, turn your speech into text, write an answer and read it back. Timing, tone and laughter get lost on the way. Full-duplex speech models keep all of that, and we want developers to be able to use them without running GPUs.
We plan to offer these models over WebRTC, phone lines (SIP) and WebSocket, with simple monthly plans and per-minute pricing.
Dessa and other models
Dessa is a full-duplex speech model announced by Block, Inc. in October 2026. It is not released yet. If Block releases it under a license that allows hosting, we plan to offer it. Either way we will host other full-duplex models. We track what is known on the Dessa release status page.
Independent site. Not affiliated with or endorsed by Block, Inc.
Where we are
The API is opening soon. You can choose a plan today and lock in your price. Nothing is charged until launch, and we will email you before billing starts.