Block's full-duplex speech-to-speech model, trained only on human conversation and currently in research preview with a public waitlist.
Dessa official page hero
Overview
What is Dessa?
Block's speech-only conversation model, learned by listening.
Dessa is a full-duplex speech-to-speech model from a small research lab at Block, the company behind Square and Cash App. According to Block, Dessa was trained from scratch on human conversation alone, with no text and no pretrained models, and it listens and talks at the same time. Block says inference runs faster than real time and the model runs locally on a MacBook Pro with an M4 Max chip or better.
As of October 2026, Dessa is a research preview, not a public product. Block has published demo recordings and a public waitlist, with early access going first to a small group of AI safety researchers. Block has not announced a public release date, an API, model weights, pricing, or benchmark results.
Platforms and languages
Dessa Availability
Platforms
Languages
Capabilities
Dessa Key Features
Speech-to-speech model
Audio in, audio out, with no text step in between, according to Block.
Full-duplex conversation
Listens and talks at the same time, so it can cut in and backchannel during a conversation.
Trained only on speech
Trained from scratch on human conversation alone, with no text and no pretrained models, according to Block.
Runs locally on a laptop
Block says inference runs faster than real time and the model runs locally on a MacBook Pro M4 Max or better.
Research preview with waitlist
Early access goes first to a small group of AI safety researchers, alongside a public waitlist.
Best for
Who uses Dessa?
AI safety researchers
test a speech-only conversation model and find risks before wider release
Voice AI builders and researchers
follow a new direction in speech-to-speech model design
Common questions
Dessa FAQs
Can I use Dessa today?
No. As of October 2026, Dessa is a research preview. Block offers a public waitlist, with early access going first to a small group of AI safety researchers.
How is Dessa different from other voice assistants?
According to Block, Dessa was trained from scratch on human conversation alone, with no text and no pretrained models. It takes audio in and gives audio out, listening and talking at the same time.
Is there a Dessa API or download?
Block has not announced a public release date, an API, model weights, or pricing. The only official route for now is the waitlist on Block's Dessa page.
What can Dessa do in the demos?
Block says Dessa learned from real conversation when to cut in and when to backchannel, and that it hears and speaks emotion, sarcasm, and laughter. These are Block's claims from its own demo recordings, not independently verified results.
Reviews
See what the community thinks and share your experience.
—
Based on 0 ratings
Rating distribution
Community reviews
No reviews
No written reviews yet.
