A New Category of Multi-Agent Voice AI on iPhone
Sesame’s iOS preview introduces a new category of conversational AI app by offering multiple lifelike voice AI agents on a single iPhone, each with its own personality, memory, and tools designed to turn spoken interactions into practical, everyday workflows that go beyond traditional single-assistant models. Instead of one default iPhone voice assistant, the Sesame app offers four named agents—Maya, Miles, Simone, and Charlie—available during a free preview across 39 markets through the App Store. Users can speak, type, or tap into these agents as “personal partners” for search, planning, and note-taking. Sesame’s design goal is to make voice-first computing feel more natural than a chatbot prompt while staying fast enough for routine tasks. This move positions the app among emerging Siri alternatives that treat voice as an always-on interface rather than an occasional utility for timers, messages, or weather checks.
Inside the Four Agents: Beyond a Single iPhone Voice Assistant
Each Sesame agent is built to sound and respond differently, giving users a choice of conversational style instead of a single default iPhone voice assistant. Maya, Miles, Simone, and Charlie are designed with distinct voices, rhythms, and personalities so that people can pick an AI that feels more like a familiar partner than a generic bot. The app keeps one continuous thread per session where users can “search, text, and think” together with an agent, blending live answers, planning, and reflection in a single conversation. Parallel search and retrieval run while you speak, pulling web results into contextual replies before the audio ends. This multi-agent approach tests whether specialized voice AI agents can handle different roles—such as planning, research, or reflection—while sharing memory and context. Sesame is betting that personality plus continuity will keep people in voice mode longer than a quick command to Siri.
Practical Tools: Search Cards, Notes, and Text Mode in One Thread
Sesame’s conversational AI app focuses on making voice useful, not only impressive. Real-time search cards layer images and key facts over spoken replies, turning a voice answer into something you can quickly scan and reuse. Within the same thread, users can ask an agent to take notes, highlight key points, or generate summaries instead of jumping to separate apps. A dedicated text mode lets people type when speaking is awkward, such as in quiet spaces, while still staying with the same agent and context. According to TestingCatalog, the agents include “comprehensive memory for individualized agent experiences,” allowing them to remember preferences and past topics across sessions unless incognito is enabled. Together, these features push voice beyond simple questions and toward everyday jobs: planning a trip, capturing ideas from a commute, or turning spontaneous questions into saved notes that stay tied to a specific AI persona.
Privacy and Latency: Can Voice Become a Daily Habit?
To compete with established Siri alternatives, Sesame has to solve two big problems: privacy worries and the awkward lag that breaks conversation flow. The app includes an incognito mode for people who want ephemeral chats that stay out of memory and off Sesame’s servers, appealing to privacy-conscious users wary of long-term data retention. On the performance side, Sesame’s parallel search and retrieval system is optimized for low first-audio latency so replies arrive quickly enough to feel human, yet detailed enough to be useful. WinBuzzer notes that the company is using this wider iPhone preview to “test whether Sesame’s voice model can build daily habits” ahead of a planned 2027 eyewear push. If users keep using voice for search, planning, and quick tasks instead of falling back to touch, Sesame will have shown that conversational AI can be a reliable everyday mode, not a short-lived demo.
Competition and the Future of Siri Alternatives
Sesame’s launch drops into a crowded field of voice AI agents that includes ElevenLabs, OpenAI’s Realtime efforts, Hume EVI 4, Vapi, and Deepgram. The race is no longer about flashy demos; it is about who can answer quickly, keep context across long chats, and sustain natural back-and-forth conversation. According to WinBuzzer, keeping first-audio latency under roughly 300 milliseconds has become a key threshold for natural exchanges. Sesame’s differentiated bet is its multi-agent structure and its roadmap toward intelligent eyewear, which could turn today’s iPhone voice assistant preview into a fuller ambient computing platform. The current free app is a proving ground: if users adopt these agents as daily companions for notes, search, and lightweight planning, Sesame will stand out among Siri alternatives. If not, it risks being remembered as a polished experiment that never escaped the novelty phase.