A New Voice AI iPhone App Built Around Conversation
Sesame’s new voice AI iPhone app is a multi-agent, voice-first assistant that turns spoken conversations into ongoing threads for search, note-taking, and planning on your phone. Instead of a single chatbot, the iOS preview introduces four conversational AI agents — Maya, Miles, Simone, and Charlie — each with its own personality and delivery style. Available through the App Store across 39 markets in a free preview phase, the Sesame voice app focuses on lifelike timing and natural back-and-forth more than on one-off prompts. The company’s work on voice presence aims to make users feel heard rather than talked at, with parallel systems balancing quick replies and more considered answers. A potential waitlist during preview helps control quality and latency as more people sign up, signaling that this iOS voice assistant is still in active development rather than a finished product.
Four Distinct Agents: Maya, Miles, Simone, and Charlie
Sesame’s conversational AI agents are framed as personal voice companions rather than generic assistants. Maya, Miles, Simone, and Charlie are tuned with different voices and personalities so users can pick a style that fits their mood or task. That multi-agent design lets Sesame test whether people prefer a consistent, almost character-like partner over a neutral system. Each agent shares access to the same underlying search and memory features, but interactions feel individualized: the app builds a profile over time, recalling details from past sessions to keep conversations coherent across days. According to TestingCatalog, the agents offer “comprehensive memory for individualized agent experiences,” while still giving users control through incognito mode when they do not want sessions stored. This mix of continuity and opt-out privacy is central to the idea that a voice companion might become something users return to many times a day.
Search Cards, Notes, and One-Thread Workflows
The Sesame voice app is structured around a single conversational thread where users can search, store notes, and get summaries without switching apps. While you speak, the agents run parallel searches and pull live web results into “search cards” — visual snippets that can include images and key facts before the audio reply ends. Notes and reminders live inside that same thread, turning a casual question into a lightweight to-do list or research log. WinBuzzer reports that the app lets people “search, text, and think” in one place, and that all-in-one flow is what Sesame is stress-testing. A text mode supports quiet or discreet use, while longer follow-up sessions keep context so you can return to the same topic later. By handling small but common phone tasks, the iOS voice assistant is aiming squarely at repeat, everyday use rather than short demo moments.
Privacy, Incognito Mode, and Memory Control
For an app that leans on long-term conversational memory, privacy controls are prominent. By default, Sesame’s agents remember past chats to personalize future replies, building continuity around preferences, ongoing projects, and recurring questions. When that persistence is not welcome, users can switch to incognito mode, which keeps conversations out of long-term memory and, according to WinBuzzer’s summary, off Sesame’s servers. This creates two clear modes: a “remember me” state for a growing relationship with an assistant, and a “forget this” state for sensitive questions or one-off tasks. The app’s design suggests Sesame understands that a true conversational AI companion needs explicit boundaries to be trusted. Coupled with text mode for more private contexts, the privacy options reinforce the idea that voice-first computing can fit into daily life without turning every spoken interaction into permanent data.
From Preview App to Everyday Companion and Future Hardware
Sesame’s iOS preview is as much a product test as a soft launch. The free release in 39 markets gives the company a wide funnel to see whether people will keep using voice for search, planning, and quick tasks once the novelty fades. Competition from players such as OpenAI Realtime, ElevenLabs, Hume EVI 4, Vapi, and Deepgram puts pressure on first-audio latency and natural turn-taking, but Sesame is betting that its multi-agent design and one-thread workflow can keep users in voice mode longer. The stakes extend beyond phones: Sesame has already signaled a roadmap toward intelligent eyewear, with a 2027 push planned after earlier hardware concepts in 2025. The question now is whether today’s conversational AI agents on iPhone can build habits strong enough to support that future, turning a preview app into the backbone of an always-available, spoken interface.