Thinking out loud
Talk through an idea, decision, outline, or question when a fluid spoken exchange is more useful than composing a formal prompt.
Independent tool overview
Sesame at a glance
Sesame is a preview-stage personal-agent experience built around expressive, low-latency voice conversation. Its mobile app offers four agents with distinct voices and personalities, plus text, web search, deep dives, notes, reminders, memory, and post-call summaries; a simpler browser Research Preview remains available for voice conversations with Maya and Miles.
Visit the official Sesame site ↗
- Product type
- Voice-first personal AI agents
- Current availability
- Mobile preview on iOS and Android; web Research Preview
- Mobile agents
- Maya, Miles, Simone, and Charlie
- Official language
- English
- Logged-in call limit
- 30 minutes per call, restartable
- Current price
- Free during the preview
- Last reviewed
- August 29, 2026
Overview
What Sesame is
Sesame began as a research demonstration of more natural conversational speech and has expanded into a mobile personal-agent preview on iOS and Android. Maya, Miles, Simone, and Charlie are intentionally different conversational characters rather than interchangeable voice skins, and each maintains its own memory across voice and text when the user is signed in.
The product is aimed at thinking aloud, exploring a topic, reflecting on a decision, and capturing ideas during moments when typing is inconvenient. Agents can search the web while speaking, produce richer Deep Dives, save notes or reminders, send a post-call summary, and show supporting search cards in the app.
Sesame is still a preview, not a dependable executive assistant or professional adviser. The company explicitly says agents can make things up and currently limits their scope to calls, text, search, deep research, notes, reminders, and summaries. Its intelligent eyewear is a separate future product scheduled for 2027, with no public price or final capability guarantee.
Use cases
Who Sesame is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Hands-free curiosity
Explore a topic while walking or commuting, with web search and later access to the supporting material in the app.
Capturing personal notes
Ask an agent to save a takeaway or reminder and use the post-call summary to preserve the thread.
Conversational voice research
Evaluate expressive prosody, timing, personality, and contextual voice behavior through the hosted preview or the separate open-source CSM 1B model.
Capabilities
Core Sesame features
Four personal agents
The mobile preview includes Maya, Miles, Simone, and Charlie, each designed with a distinct voice, temperament, and conversational point of view.
Voice and text
Move between spoken calls and written conversation while keeping the selected agent's identity and memory.
Live web search
Agents can run multiple searches while speaking and incorporate new results into the answer, with visual search cards available in the app.
Deep Dives
Creates a more in-depth result when a quick conversational response is not enough for the topic.
Notes and reminders
Saves useful ideas or follow-ups from the conversation so they can be revisited later.
Post-call summaries
Provides a written recap after a call rather than requiring the user to remember or transcribe the whole exchange.
Agent-specific memory
Each character remembers its own prior conversations, and memory carries between web and mobile under the same signed-in account.
Incognito Mode
Lets an agent use prior memory while preventing the current conversation from being saved as new memory, according to Sesame's product guide.
Open-source CSM 1B
Sesame separately publishes an Apache-licensed speech-generation model for research. It generates audio from text and audio context but is not the full personal-agent product or a general-purpose language model.
Process
How the Sesame workflow works
Step 1
Choose the right character
Preview the agents and pick the conversational style that fits the task; their memories are separate, so use one consistently for a continuing topic.
Step 2
State the purpose aloud
Tell the agent whether you want brainstorming, a decision framework, factual research, reflection, or a saved action item.
Step 3
Ask for sources when facts matter
Use search cards or a Deep Dive, then open the underlying sources instead of treating a confident spoken answer as verified.
Step 4
Capture the useful part
Save a note or reminder during the conversation and review the post-call summary while the context is still fresh.
Step 5
Use Incognito deliberately
Enable Incognito before a conversation you do not want added to memory, while remembering that legal privacy terms—not the conversational agent—control data handling.
Step 6
Escalate high-stakes decisions
Take medical, mental-health, financial, legal, safety, and other consequential questions to a qualified human professional and independently verify every factual claim.
Cost
Sesame pricing and free plan
Sesame says the full mobile preview is currently free. No paid consumer tiers, usage prices, enterprise rate card, or eyewear price are publicly listed as of August 29, 2026. Logged-in mobile and web calls are capped at 30 minutes and can be restarted immediately; the web Research Preview allows 5-minute calls without logging in. The separate CSM 1B code and checkpoint are available for research under Apache 2.0, but self-hosting requires compatible hardware and carries its own compute and engineering cost.
Mobile Preview
Free during preview
The fullest current Sesame product on iOS and Android.
- Voice and text
- Four agents
- Search, Deep Dives, notes, reminders, and summaries
- 30-minute calls when signed in
- Preview terms and limits can change
Web Research Preview
Free preview
A simpler browser experience focused on voice calls with Maya and Miles.
- Voice calls and web search
- 5-minute sessions without an account
- 30-minute sessions when logged in
- Fewer features than the mobile preview
CSM 1B
Free and open source
A separate Apache-licensed conversational speech-generation model for technical research.
- Requires the CSM and Llama 3.2 1B model access
- CUDA-compatible GPU recommended
- Generates speech rather than text reasoning
- Cloud or local compute is not free
Sesame eyewear
Not announced
Future hands-free hardware targeted for 2027.
- No public purchase price
- No final availability or production specification
- Do not buy the app preview assuming hardware terms are fixed
Pricing checked . Check current pricing at the source ↗
Assessment
Sesame strengths and limitations
Where it stands out
- Expressive speech, timing, and distinct agent personalities make voice interaction feel less mechanical
- Voice, text, search, Deep Dives, notes, reminders, and summaries form a coherent mobile workflow
- Agent-specific memory supports continuity without blending every character's history
- Incognito Mode provides a visible control for conversations users do not want added to memory
- The current preview is free on mobile and web
- Sesame has published core speech-generation research and an open-source 1B model
What to consider
- The service is a preview and can change, fail, or impose new availability and pricing conditions
- Agents can fabricate facts and are not substitutes for medical, therapy, legal, or financial professionals
- English is the only officially supported language
- Calls are capped at 30 minutes when logged in and 5 minutes for anonymous web use
- Logged-in content may be retained and used for legitimate business purposes, including model training, under Sesame's current terms and privacy policy
- The open-source CSM 1B model is speech-only, primarily English, hardware-intensive, and not the same system as the hosted personal agents
- Human-like voice and personality can encourage overtrust or emotional anthropomorphism despite the system's factual limits
Compare
Sesame alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Agents
11ai
Consider ElevenLabs 11ai when voice-agent integrations and developer deployment matter more than a consumer personal-agent experience.
Explore 11ai →Sales
Vapi
Consider Vapi for building customizable business voice agents across phone, support, sales, and service workflows.
Explore Vapi →Business Operations
ChatGPT
Consider ChatGPT for a broader multimodal assistant with mature tools and voice across general productivity use cases.
Explore ChatGPT →Content Creator
ElevenLabs Text To Speech GPT
Consider the ElevenLabs text-to-speech workflow when the goal is producing voiced content rather than maintaining an ongoing personal-agent relationship.
Explore ElevenLabs Text To Speech GPT →Questions
Sesame FAQs
What is Sesame AI?
Sesame is a voice-first personal-agent preview built around natural conversational timing, expressive speech, distinct personalities, search, memory, notes, reminders, and summaries.
Is Sesame free?
Yes, Sesame says its current mobile and web previews are free. It has not announced permanent paid tiers or pricing for its planned 2027 eyewear.
Which Sesame agents are available?
The mobile preview offers Maya, Miles, Simone, and Charlie. The web Research Preview is centered on voice conversations with Maya and Miles.
Does Sesame work on Android?
Yes. Sesame's current Mobile Preview page says the preview is available on iOS and Android, superseding its May 2026 launch post that described Android as forthcoming.
How long can a Sesame call last?
Signed-in calls are capped at 30 minutes and can be restarted immediately. Anonymous web sessions are limited to 5 minutes.
Does Sesame save my conversations?
Signed-in agents use past conversations for memory, and Sesame's current terms allow logged-in content to be retained and used for business purposes including model training. Sesame says Incognito Mode does not save the current conversation as new memory. Review the live privacy terms before discussing sensitive material.
Is Sesame CSM the same as the app?
No. CSM 1B is an open-source speech-generation model that produces audio from text and audio context. It does not generate text or provide the full agent, search, memory, and mobile experience.
Bottom line
Our Sesame verdict
Sesame is one of the more distinctive voice-agent previews because it treats timing, prosody, personality, search, memory, and capture as one conversational product. It is best used as a curiosity and reflection tool, not an authority. The preview is easy to try for free, but its human-like delivery raises the need for stricter fact-checking and privacy discipline—not less. Use Incognito when appropriate, keep sensitive and regulated data out, and verify any consequential answer outside the conversation.
Visit Sesame website ↗Similar Content Creator Tools
View all Content Creator AI tools →
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.