Voice & AI audio
Sesame's talking-AI app, explained — and is it worth trying?
A new iPhone app lets you chat with a remarkably human-sounding AI. Here's the plain-English rundown.
The answer
Sesame's iPhone app, launched 28 May 2026, lets you talk to a natural-sounding AI assistant.
If you've tried talking to Siri, Google Assistant or a standard voice assistant and found it stilted or robotic — pausing awkwardly, misunderstanding your tone, or just sounding flat — Sesame is trying to fix exactly that. Here's what its new iPhone app is, how it works, and whether it's actually worth downloading.
Who is Sesame and why should you trust them?
Sesame is a US startup co-founded by Brendan Iribe (a co-founder and former CEO of Oculus — the virtual-reality company Meta, formerly Facebook, acquired in 2014) and Ankit Kumar, with other Oculus veterans on the team. That background is relevant because building convincing, low-latency audio experiences is hard: it's the same core challenge as making VR feel immersive. The team has been working on this particular problem for years, not months — and it's well-funded, having raised a $250 million round led by the investment firm Sequoia before the app launched.
Before they launched the app, Sesame published the underlying technology — components of their Conversational Speech Model, including the 1B 'CSM-1B' backbone — as open-source software under a licence that lets anyone inspect and use it (Apache 2.0). That meant developers and researchers could run and test it publicly before any consumer product launched. So when Sesame says its voices sound human, the model behind that claim is out in the open for anyone to check — not a black box you have to take on trust. That's rarer than it sounds in AI.
Sesame, the conversational AI startup from Oculus founders, launched its iOS app on 28 May 2026 — a public preview in 39 countries with four AI agents, built on its Conversational Speech Model.
What makes Sesame's voices actually sound different
Most AI voices are built to be accurate and clear. Sesame's is built to be natural. There's a difference. Here's a plain-English breakdown:
| Feature | Standard AI voice | Sesame's CSM |
|---|---|---|
| Pauses | Fixed, mechanical timing | Natural rise-and-fall rhythm |
| Tone | Flat, constant delivery | Matches the emotional register of what it's saying |
| During 'thinking' | Freezes mid-conversation | Designed to keep the dialogue flowing |
| Feels like | A screen reader | A phone call with a person |
The technical name for this is prosody — the music of speech. Sesame built their model specifically around getting prosody right, rather than bolting voice on top of a text-based AI after the fact. That's a real engineering choice that makes a real audible difference.
The app gives you four different AI 'agents' to choose from — Maya, Miles, Simone and Charlie — essentially four distinct voice personalities, each with its own character and its own memory of your past chats. Whether you want something more brisk and professional or warmer and conversational, you can pick. That's a thoughtful design decision: the voice you want while driving is probably not the same one you want while cooking. There's also a clever trick under the hood — instead of going silent while it 'thinks', the AI can run searches while it's still talking and fold the results into what it's saying, so the conversation keeps flowing.
Should you try it?
If the idea of chatting with an AI rather than typing appeals to you — while cooking, walking, driving, or just when your hands are busy — Sesame is the best place to try that right now. The voice quality is genuinely impressive, and unlike many early-stage apps, this one has credible engineering behind it. It's available on iPhone in 39 countries as a public preview.
Sesame launched its iPhone voice AI app with four agents — Maya, Miles, Simone and Charlie — in 39 countries, with live search, notes, summaries and an incognito mode all in one spoken session.
How does it compare to Siri, ChatGPT voice and Google Assistant?
These are the main alternatives most people already have on their phone:
| Sesame | ChatGPT Advanced Voice | Siri AI | Google Gemini voice | |
|---|---|---|---|---|
| Voice naturalness | Best-in-class | Very good | Improved (WWDC 2026) | Good |
| Where it lives | Separate app | ChatGPT app | Built into iPhone OS | Google app |
| Other AI features | Voice only | Full ChatGPT | Deep iPhone integration | Full Gemini suite |
| Cost | Free preview | Free / Plus tier | Free with iPhone | Free |
Sesame wins on voice naturalness specifically. If that's what matters to you — having an AI that's genuinely pleasant to talk to — it's the obvious first choice for a try. If you want an AI that can also look up your emails, set calendar reminders, and control your apps, the native options have the advantage.
Where to find it: it's available on the App Store as a public preview. Being a preview means features and availability can change, and the free access may eventually give way to paid tiers — standard practice for early-access launches. Check the listing for current terms before you dive in deeply.
Frequently asked questions
How do I get Sesame's app?
Is Sesame's app free?
Is it available on Android?
What are the four AI agents?
Is Sesame's voice AI safe to use?
Sources
- Sesame, the conversational AI startup from Oculus founders, launches its iOS app — TechCrunch, 28 May 2026
- Sesame Launches iPhone Voice AI App with Four Agents — WinBuzzer, 29 May 2026
- Crossing the uncanny valley of conversational voice (CSM; models under Apache 2.0) — Sesame, 27 February 2025