Developers
Dessa AI API Use Cases: What Developers Could Build If It Opens
Search interest in a Dessa API keeps earning clicks on this site, so it is worth answering the next question developers ask carefully: if Block ever opens an API, what could you build with it? The starting fact has not changed. There is no Dessa AI API today — no endpoints, no documentation, no SDK and no keys. Block has not announced one. Everything below separates that record from the use cases that follow logically from what Dessa is verified to do, so you can prepare without planning around a product that does not exist.
First, what exists — and what does not
Our original Dessa AI API status article records what Block has and has not announced, and our documentation, SDKs and keys article checks the concrete artifacts developers look for. The answer in both cases is the same: none of them exist publicly. Our developer preparation guide is therefore about readiness — architecture choices you can make now — not integration steps, because there is nothing to integrate yet.
Dessa itself is a research preview. Block announced it on October 23, 2026 as a full-duplex, speech-to-speech model trained from scratch on human conversation audio only, with no text and no pretrained models. Early access goes to a small group of AI safety researchers first. A public waitlist sits at block.xyz/dessa. No API, pricing for an API, or release date has been announced — our API pricing article covers that blank separately.
Use cases that follow from verified abilities
Dessa's verified abilities are conversational, not clerical. Block's demos show a model that listens while it speaks, hears emotion, sarcasm and laughter, backchannels naturally, adjusts volume when asked, and knows when to interrupt or let a sentence finish. If an API ever exposed that kind of live conversation, the natural use cases would be the ones where how something is said matters as much as what is said.
Customer support is the clearest example. Our article on Dessa AI for customer support explores what changes when a system can hear frustration instead of only transcribing words. Voice ordering for restaurants is another — we covered the Square angle in Dessa AI and Square restaurants, clearly labelled as analysis, not a Block product announcement. Spoken practice and tutoring follow from the same abilities, as discussed in our education and language learning article, with the caveat that Block has published no supported-language list.
Use cases Dessa is not verified for
Just as important is what not to assume. Dessa is not text-to-speech, as our TTS explainer sets out — it does not start from written text. Block has not presented it as a transcription service, a tool-calling agent, a reading system or a factual database, and our article on what Dessa can't do yet covers those limits. An API use case that needs Dessa to browse, book, calculate or reliably retrieve facts is not supported by anything Block has published. Planning such a product today would be planning around invented features.
Accessibility is another honest example. A conversation system that hears tone and responds in speech, in real time, could matter to people who find typing or reading screens difficult — if it were ever offered that way. Block has not announced an accessibility product, a mobile app or phone support for Dessa, so that too belongs in the preparation column, not the product column. The pattern is consistent: the more a proposed use case depends on a specific Block announcement, the less of it you should build today.
What to do while there is no Dessa API
Developers who need live voice today do have options that actually exist. Our Dessa API vs OpenAI Realtime API comparison makes the contrast plain: one API is available, the other is not. Our alternatives you can use today article covers the wider field. The sensible preparation is to build your voice experience so the conversation layer can be swapped later, keep your own handling of consent, recording and safety, and watch the official page — not rumours — for any change in Dessa's status.
If Block announces an API, this site's Tracker will change and we will cover the documentation on the day it exists. Until then, the honest developer summary is: interesting use cases, no API to build them on yet.
Sources: Block's official Dessa page at block.xyz/dessa, checked November 14, 2026. No API, API pricing, documentation, SDK, weights or release date had been announced as of that check.
Frequently asked questions
Can developers build on a Dessa AI API today?
No. Block has not announced or released a Dessa API, documentation, SDK or API keys, so there is nothing to build on today. This article discusses preparation and possible use cases only.
What could a Dessa API be used for if Block releases one?
The use cases follow from Dessa's verified abilities — live, emotional, full-duplex conversation — such as voice support, voice ordering and spoken tutoring. Block has not promised any of these as API features.
What should developers use instead while waiting?
Available realtime voice APIs exist today, including OpenAI's Realtime API. Our API comparison and alternatives articles cover what you can actually use now, with Dessa's status stated honestly.
🚀 Get Dessa launch alerts
Dessa is still a research preview. Join the free DessaAI Daily list and we will email you the day access opens — plus the week's biggest Dessa stories, once a week. No spam, no hype.
Form not loading? Subscribe on the signup page →