# Block announcement of Dessa voice AI model

> Live situation record from CLSTR: https://clstr.news/situations/block-announcement-of-dessa-voice-ai-model
> Updated: 2026-10-26T10:40:00.000Z. Sources: 33. Developments: 2.

Block, the fintech company led by Jack Dorsey, has introduced Dessa, a research preview of a new full-duplex, speech-to-speech AI model. Unlike traditional assistants that convert speech to text, Dessa is trained exclusively on raw human audio to capture emotional nuances such as sarcasm, laughter, and sighs.

The model is designed to listen and speak simultaneously, allowing for natural interruptions and the use of fillers. Technical specifications indicate the model can run locally on consumer hardware, such as a MacBook Pro with an M4 Max chip. 

To address potential risks, including the use of highly realistic voices for fraudulent activities, Block is currently providing early access to a small group of AI safety researchers rather than the general public.

## Timeline

### 2026-10-26: Block introduces Dessa, a human-like full-duplex voice AI model

Block has launched Dessa, a full-duplex voice AI model that runs locally and features human-like emotions, pauses, and natural speech patterns.

2 sources. https://clstr.news/cluster/block-introduces-dessa-a-human-like-full-duplex-voice-ai-model-1

### 2026-10-23: Block unveils Dessa, a speech-to-speech AI trained only on human audio

Block has introduced Dessa, a full-duplex voice AI trained solely on human speech. It processes audio directly to preserve emotion and can run locally on high-end laptops like the MacBook Pro M4 Max.

31 sources. https://clstr.news/cluster/block-unveils-dessa-an-emotional-ai-voice-model

---
Cite as: Block announcement of Dessa voice AI model. CLSTR, https://clstr.news/situations/block-announcement-of-dessa-voice-ai-model