Sesame AI Launches iOS App for Natural Conversational Agents
Sesame, the conversational AI startup co-founded by Oculus veterans, has launched its iOS app, bringing voice-forward AI agents designed to mimic natural human conversation to the public. The app targets the gap between stiff chatbot interactions and genuinely fluid back-and-forth dialogue.
Original sourceSesame, founded by Brendan Iribe and Michael Antonov — two of the key figures behind Oculus — has officially launched its iOS app after a period of limited access. The product centers on conversational AI agents built to handle the interruptions, context shifts, and natural cadence of real human conversation, rather than the turn-based ping-pong of most chat interfaces.
The core technical bet is on low-latency, high-context voice interaction. Sesame has been vocal about the idea that the uncanny valley in AI conversation is less about intelligence and more about timing — the awkward pauses, the inability to be interrupted, the loss of context mid-sentence. The app attempts to close that gap with what the team describes as 'presence,' a design principle borrowed from immersive media, which makes sense given the founders' backgrounds in VR.
The iOS launch marks Sesame's first major consumer-facing move after months of demos and waitlist-gated previews. The app is free to download, with a subscription tier for extended usage. The company has not disclosed specifics on model architecture or third-party API dependencies, which leaves questions about the underlying stack and whether its conversational quality holds up outside controlled demo conditions.
For context, the conversational AI space is crowded — OpenAI's Advanced Voice Mode, Hume AI's empathic voice interface, and a growing number of voice-layer startups are all competing for the same use case. Sesame's differentiation claim is feel rather than capability, which is harder to benchmark and harder to defend, but also harder for a competitor to copy quickly if the team has genuinely cracked something in the interaction model.
Panel Takes
The Skeptic
Reality Check
“The direct competitor here is OpenAI Advanced Voice Mode, which ships to hundreds of millions of users on day one — and Sesame is asking people to download a separate app for what is, functionally, a voice interface layer. The 'presence' framing is interesting but untestable without methodology, and 'feels more human' is exactly the claim every voice AI startup makes right before the demo collapses on real-world ambient noise and topic switches. What kills this in 12 months: OpenAI and Google ship marginally better latency natively and the distribution moat evaporates overnight.”
The Founder
Business & Market
“The buyer here is unclear — is this B2C subscription, or is the iOS app a trojan horse for an enterprise or developer API play? Iribe and Antonov have serious credibility and presumably the fundraising to match, but 'presence' is a brand attribute, not a moat, and a subscription for a voice chat app has brutal churn dynamics unless the use case is sticky enough to be habitual. The real question is whether the iOS launch is the product or the demo reel for an SDK that's coming next — because the business that survives is the one selling the voice layer to other apps, not competing for home screen space against ChatGPT.”
The Creator
Content & Design
“The output here is conversation itself, which means the taste layer lives entirely in pacing, word choice, and how the agent handles the moment you go off-script — and without a public demo I can actually run, I'm evaluating a promise, not a product. What I'll say is that the VR founders framing is legitimately relevant: immersive media taught people that presence is an emergent quality of small decisions made consistently, not a single feature, and if that instinct transferred to voice design the output could feel genuinely different from the ChatGPT voice cadence. But 'feels more human' as shipped output is going to look like every other AI assistant the moment someone asks it something weird.”
The Futurist
Big Picture
“The thesis here is falsifiable: in three years, the dominant interface for AI is ambient voice, and the teams that win are the ones who optimized for conversation feel rather than raw capability — meaning latency, interruption handling, and context persistence matter more than benchmark scores. That's a plausible bet, and the Oculus founders are credible people to make it given they spent years solving the adjacent problem of perceptual presence in VR. The dependency that has to hold: voice AI remains a distinct enough interaction modality that it doesn't get commoditized into a settings toggle inside existing apps — and right now that dependency is fragile.”