Talk to your Agent the way you would talk to a collaborator sitting next to you. Live voice is now available in Picsart Agents, built on GPT-Live-1, the real-time voice model OpenAI released today.
Until now, working with an Agent meant typing. You wrote a brief, waited, read what came back, then typed again. That works, but it is not how anyone actually talks through an idea. You interrupt. You change your mind halfway through a sentence. You point at something and say “not that, the other one.”
Live voice closes that gap. You speak, the Agent responds as you go, and neither of you has to wait for the other to finish.
The reason it feels different comes down to how the model works. Most voice assistants pass your words through a chain: one system turns speech into text, another works out a reply, a third reads it back. Every handoff costs time, and it is why talking to them feels like waiting your turn. GPT-Live-1 listens and speaks at once, in a single model, so it can react while you are still talking.
Speak and hear back in real time
There is no send button and no pause while a full response is assembled. You talk, the Agent talks, and you can jump back in the moment you have something to add. It is the difference between messaging someone and being on a call with them.
Change direction without starting over
Halfway through a task you can add context, correct yourself, or send the Agent somewhere else entirely. Interrupt mid-sentence and it adjusts. Nothing resets, and you do not lose the work already done.
Create hands-free on mobile
Describe your next flow, ad set, or edit out loud on iOS without typing a prompt. Useful when your hands are busy, when you are away from a desk, or when it is simply faster to say the thing than to write it.
Every Agent, one conversation
Voice works across the whole Agent roster, so it is the same conversation whichever specialist you are working with. Ask Mira to shape a brand identity, walk Reeva through the reel you have in mind, or talk Theo into the right background for a product shot. If you are not sure who you need, Aura will route the brief to whichever specialist fits.
Where it works today
Live voice is available now in Picsart AI Agents on web and iOS, and it will reach more of Picsart shortly. You do not need to set anything up: open an Agent, start talking.
GPT-Live-1 launched in the OpenAI API today, and Picsart is one of the products featured in the announcement alongside Yelp, Cognition and HeyGen. Here is how OpenAI describe what we built with it:
“Powered by OpenAI’s GPT-Live-1, Picsart’s creative AI Agents let creators brainstorm, generate, refine, and publish content through real-time voice conversation.”
You can read the full announcement on the OpenAI blog.
