Skip to content
All posts

Pricing8 min read

Sesame AI pricing 2027: is the Maya voice free?

Sesame AI is free in a research preview with no paid plan and no published pricing as of October 2026. Sessions cap at about 30 minutes, and its CSM voice model is open-weight under Apache 2.0, so developers can self-host it rather than buy a Sesame API.

Sesame AI pricing: a developer at a desk using voice AI tools for text to speech, speech to text and AI phone calls

App

$0free

research preview

iOS and Android

Session

30 minlimit

per session

5 min non logged in

CSM-1B

Openweight

Apache 2.0

self-host the model

Host rate

$7/1M chars

via OpenRouter

host rate, not Sesame

Sesame pitches conversational voice agents and AI eyewear, with paid tiers signaled but not published. The $7 per million characters figure is a third-party host rate for the open model, not an official Sesame price.

Key takeaways

  • Sesame AI costs $0 today: the Maya and Miles app is a research preview with no paid plan and no published rate card.
  • Preview sessions stop near 30 minutes, and visitors who skip login on the web get roughly 5 minutes per call.
  • CSM-1B, the open Sesame speech model, ships under Apache 2.0, and OpenRouter hosts it near $7.00 per million characters.
  • For billed production voice, PyAI Speak runs $0.04/min and Omni $0.05/min + $0.01 telephony, with $50 of starter credit and no credit card.

Sesame AI plans compared: free preview and open model

Sesame AI is a conversational voice research company whose product is currently a free preview, with an open-weighted speech model. The rows below set out where Sesame pricing stands, with third-party figures flagged.

Research preview
$0, free

The consumer app on iOS and Android is a free research preview, with no paid plan and no published minute allowance.

Session limits
About 30 minutes

Sessions cap at roughly 30 minutes, with about 5 minutes a call for non-logged-in web users, as a preview rather than a service.

CSM-1B model
Open weight, Apache 2.0

The Conversational Speech Model behind the Maya and Miles demos is an open-weight model, so developers can self-host it for free aside from compute.

Third-party host
About $7 per million characters

OpenRouter lists CSM-1B at roughly $7 per million characters, a host rate for running the open model, not Sesame's own price.

Future paid tiers
Signaled, not published

Sesame has signaled paid tiers possibly tied to AI eyewear or an API, but nothing is published, so there is no commercial rate yet.

How much Sesame AI really costs

Sesame costs nothing to use in preview, since it is free with no paid plan. The only real cost today is compute if you self-host the open CSM model, or a third-party host rate for running it.

App$0, free previewNo paid plan, no published allowance
Session limitRoughly half an hour5 minutes for non-logged-in web
Open modelSelf-host CSM-1BApache 2.0, you pay compute
Host rateAbout $7 per million charsOpenRouter, host rate not Sesame

Which Sesame AI plan fits by use case?

Sesame today fits people exploring realistic conversational voice rather than teams deploying production. The notes below weigh who it suits and when a hosted production service is the better choice.

Curious users
Trying the demo

Anyone who wants to experience the Maya and Miles conversational voice for free in the preview app.

Researchers and developers
The open model

Developers who want to self-host the open CSM model for experiments, aside from compute cost.

Builders needing production
A hosted service

Teams that need a production voice API with pricing and an SLA should use a hosted service, since Sesame offers neither yet.

Phone and voice agents
Per-minute billing

Teams building phone agents need a billed, hosted service rather than a free preview with session limits.

Treat Sesame as a preview of realistic conversational voice. For a production deployment with pricing, a hosted service or self-hosting the open model are the real options.

Sesame AI vs PyAI: free preview vs production per-minute

Sesame preview
PyAI Speak

Sesame is free in preview with no production billing, so its cost is $0 but it is not a production service. PyAI figure = minutes times the per minute rate, a hosted, billed service. Omni adds $0.01 telephony to $0.05/min.

Sesame AI hidden costs: limits, compute, and host rates

  • No commercial pricing: Sesame publishes no rate card and has no paid plan, so you cannot budget an official cost.
  • Session limits: about 30 minutes a session, and roughly 5 minutes a call for non-logged-in web users, cap the free preview.
  • No production SLA: as a research preview there is no uptime commitment, so it is not suited to mission-critical use.
  • Open model compute: self-hosting CSM-1B is free of license fees but you pay for the compute to run it.
  • Host rates vary: the $7 per million characters figure is one host's rate for the open model, not Sesame's, so it is not an official price.
  • Future tiers unknown: paid plans are signaled but unpublished, and Sesame has published no future plan prices.

Compare Sesame on voice quality and session limits, since it sells no production API to price against.

Sesame AI reviews: what users say

We looked for real, attributed user reviews of Sesame AI that mention pricing and the product. As of October 2026 we found no public user reviews on the sources below, so there is nothing to quote honestly.

No public user reviews found

Evaluate Sesame AI on a short trial and a reference call rather than on ratings, since no verifiable user reviews exist yet.

Where we lookedG2 (0 reviews) · Capterra (0 reviews)

Sesame AI vs PyAI: preview versus production

The core difference is a free research preview versus a production service. Sesame is a free app with an open model and no pricing. PyAI is hosted, billed voice infrastructure with published per-minute rates. The table lines them up.

Sesame AIPyAI
StatusFree research preview✓ Production, supported
Pricing modelFree, no paid plan✓ Published, per minute
Production APINo public API✓ OpenAI-compatible Hear API
Free to startFree preview, no paid plan✓ $50 credit, no credit card
Realistic conversational voice✓ Maya and Miles demos, open CSMSpeak and Cast voices, plus Omni agents
Session limitsCapped near 30 minutes a session✓ Metered per minute of audio
Output for phone systemsNot a phone platform✓ PyAI supplies WAV, PCM or G.711 for phone lines
Paid speech to text and TTS APIsNo✓ PyAI Hear, Speak, and Cast at published minute rates
Open model to self-host✓ CSM-1B, Apache 2.0No, PyAI is hosted
Hosted serviceNo✓ Yes, billed per minute
Sesame AIA free preview of realistic voice

Choose Sesame if you want

  • To try strikingly realistic conversational voice in a free public demo before any build work.
  • An open model you can self-host and experiment with on your own GPU, free of usage fees.
  • A way to run the open CSM model yourself under Apache 2.0, with no vendor contract, if you already have GPU capacity.
Cost$0 in preview, no paid plan
Visit the Sesame site →
PyAIVoice infrastructure, priced by the minute

Choose PyAI if you want

  • A hosted, billed voice service you can run in production today.
  • Published per-minute pricing, billed on the minutes you actually use.
  • Speech to text, text to speech, and a phone agent on one bill, with each line metered separately so you can see the split.
Starts at$0.04/min Speak, $0.05/min Omni
Get a sandbox key →

Sesame suits anyone who wants to hear realistic conversational voice at no cost, or to self-host the open model for research. PyAI suits teams that need production voice with published pricing, no session limits, and phone-ready audio. The two sit at different stages: a free preview versus hosted, billed voice infrastructure.

Sesame AI cost by volume: free preview versus production

Sesame is free in preview, so there is no usage cost to scale, but it is not a production service. The table contrasts Sesame with a hosted per-minute service for teams weighing production.

StageSesameProduction pathNote
Trying it outFree previewFree previewSession limits apply
ExperimentingSelf-host CSM-1BCompute costOpen model, engineering work
ProductionNo service yetPyAI Speak at $0.04/minBilled per minute

Sesame free preview versus PyAI production per-minute

Sesame is $0 in preview but not production. PyAI Speak is a published $0.04 a minute as a hosted service. The comparison is about readiness for production, not undercutting a free demo.

Minutes a monthSesameSesame costPyAI SpeakPyAI Omni
60Preview$0.00$2.40$3.60
300Preview$0.00$12.00$18.00
600Preview$0.00$24.00$36.00
1,200Preview$0.00$48.00$72.00
3,000Preview$0.00$120.00$180.00

Sesame is free in preview with session limits and no production API, so $0 does not mean a production service. PyAI figures are published hosted rates, and the Omni column adds $0.01 telephony to $0.05/min.

Sesame AI pricing: the bottom line

Sesame charges nothing for its preview app as of October 2026, sells no plan, and ends each session near the 30-minute mark. CSM-1B is an open-weight voice model under Apache 2.0, so developers can self-host it, and a third-party host lists it near $7 per million characters. For production voice, PyAI publishes $0.04/min for Speak and $0.05/min + $0.01 telephony for Omni, with $50 of free credit and no credit card.

PyAI is not a free consumer preview and does not aim to match Sesame's demo experience. PyAI is production voice infrastructure priced per minute, with speech to text, text to speech, and a phone agent on one bill.

Sesame AI pricing FAQs

The questions people ask most about Sesame AI pricing.

Is Sesame AI free?

Yes. As of October 2026 the Sesame app costs nothing to use, and the company lists neither a paid plan nor a minute allowance. Sessions are limited to about 30 minutes, and non-logged-in web users get about 5 minutes a call. Sesame runs no public pricing page, so no paid tier exists to buy yet.

Does Sesame AI have an API or pricing?

Sesame does not publish API pricing. The consumer app is a free preview, and the company has signaled future paid tiers, tied to its AI eyewear or an API. For now there is no public rate, so you cannot budget a Sesame API cost from an official figure.

What is Sesame's CSM model?

Sesame's Conversational Speech Model powers its viral Maya and Miles voice demos. CSM-1B is an open-weight model under Apache 2.0, so developers can self-host it or run it through third-party hosts. One host, OpenRouter, lists CSM-1B at about $7 per million characters, which is a host rate, not Sesame's price.

Can I use Sesame AI for business calls?

Not directly today. Sesame is a free consumer research preview without a production API or published pricing, so it is not set up for business phone calls. The open CSM model can be self-hosted, but running it in production is engineering work rather than a hosted service.

What are Sesame AI's limits?

The preview caps sessions at about 30 minutes, with roughly 5 minutes a call for non-logged-in web users, and there is no commercial pricing or SLA. Sesame offers a demo of realistic conversational voice rather than a production service, so treat the app as a preview.

What is a production alternative to Sesame AI?

For production voice with published pricing, PyAI is per minute: Speak text to speech at $0.04/min, Omni at $0.05/min + $0.01 telephony, with $50 of free credit and no credit card. If you want the open CSM model, you can self-host it, while PyAI gives a hosted, billed service.

More Sesame AI comparisons and guides

Price PyAI against your own volume.

Per-minute pricing with no plan fee. Speak is $0.04/min, Cast $0.02/min, Hear $0.001/min, and Omni $0.05/min + $0.01 telephony, with $50 in starter credit and no credit card.