LISTED PROFILE

Sesame

Sesame is a preview-stage collection of personal agents centered on low-latency voice and text conversation. Users choose among four named agents — Maya, Miles, Simone, and Charlie — each with a distinct personality and memory that persists across voice and text interactions. Sessions can include web search, notes, reminders, and post-call summaries, and recent builds add image and voice-memo sharing. The mobile preview is free, with a gradual regional rollout.

Last checked2026-09-06

Product image for Sesame

01 — OUR TAKE

Who and what it suits

Per its official website, Sesame presents itself as a collection of personal agents rather than a task-first assistant, with intelligent eyewear announced for 2027. Its differentiator is low-latency voice conversation with named, persistent personas; the App Store listing confirms an 18+ rating and English-only support. Its privacy policy, last edited July 23, 2026, allows model training on logged-in content unless a user opts out, while App Store privacy labels remain developer-reported. Some reviews repeatedly praise voice quality.

Best for

  • English-speaking users who want a low-latency voice agent for thinking out loud and brainstorming
  • People who want web search, note-taking, and reminders inside a conversational agent
  • Users drawn to named personas like Maya or Miles with memory that persists across chats

Not ideal for

  • Users who need official language support beyond English
  • People who want a finished product rather than a region-gated preview
  • iPhone users on iOS 17 or earlier, who cannot install the current app

What stands out

  • Low-latency voice conversation with four named personalities
  • Web search, notes, reminders, and post-call summaries in one chat flow
  • Memory shared across voice and text conversations
  • Image and voice-memo sharing in recent app builds

What to consider

  • English is the only officially supported language
  • The mobile product is an 18+ rated preview with a gradual regional rollout
  • Logged-in content may be used for model training unless the user opts out
  • App Store privacy labels are developer-reported and not verified by Apple