Skip to content
New · the open voice benchmark is liveRead it
cantari
The shape of free, printed

Free text to speech that says what free means.

Most free text to speech pages keep the catch below the fold. Ours fits on a receipt: about 10 premium minutes a month, unlimited generations on the open-weight Kokoro engine, and no card at the door.

Start freeSee the full tool →
Free plan, itemized$0 forever
Premium engine minutes
about 10 / mo
Metered as
10,000 characters
Kokoro generations
Unlimited
Card to start
None

Converted by the house math: about 1,000 characters is about a minute of speech. The cap above is the one the generation route enforces, read from the same file.

Try a cue

Same line. Three deliveries.

The bracketed cue is an instruction the engine acts. Pick one and press play: each take is real Gemini Flash output, recorded unedited.

Pick a mood

[whispering]Don't go in there... something moved.

Real Gemini Flash output, unedited. About 1,000 characters is a minute of audio.

The catch, in full

Every limit we have, in one table.

Pages built on the word free tend to save this part for the checkout screen. Here it is at the top instead, next to exactly what paying adds.

WhatFreeCreatorStudio
Monthly price$0, forever$15 / mo$49 / mo
Kokoro engine (open weight)Unlimited, unmeteredUnlimited, unmeteredUnlimited, unmetered
Premium engine timeAbout 10 minutes (10,000 characters)About 10 hours (600,000 characters)About 30 hours (1,800,000 characters)
At the premium capPremium waits for the monthly reset; Kokoro keeps goingSame rule, a much bigger capSame rule, the biggest cap
Cloned voices you ownNot included1 owned cloned voice, exportable (coming soon)5 owned cloned voices (coming soon)

The numbers come from the plan registry the billing code reads, not from a marketing document. The full cards live on pricing, and the unmetered lane has a page of its own.

How free stays free

An open-weight engine pays for the free lane.

Kokoro is an open-weight model: its weights are published, anyone can host it, and hosts competing on identical output have pushed the cost of serving a draft close to a rounding error. Counting those generations would buy us nothing, so we do not count them, on any plan.

The expressive premium engines bill real money for every character they speak, and that is precisely the part we meter. Your free premium minutes have a printed edge rather than an invisible one, and an upgrade buys more of the expensive thing, not permission to keep using the cheap one. That split is the entire economics of free text to speech here, and it is why the free lane has no expiry date.

The whole calculation, raw engine rates included, is written out in why our free tier has an unlimited engine.

Asked before signing up.

The questions a free tier should answer plainly, with the numbers it actually enforces.

Is it really free, or a trial in disguise?
A plan, not a trial. Free text to speech here means the $0 tier renews itself every month with the same printed allowance: about 10 premium minutes plus unmetered Kokoro generations. Nothing expires, nothing converts to paid behind your back, and there is no countdown waiting at the end of week two.
Do I need a credit card to start?
No. Signing up asks for an email, not a payment method. A card enters the picture only on the day you choose Creator or Studio, and both prices are printed before you click anything.
What happens when I hit the monthly limit?
The premium engines pause until the month resets, and the studio warns you before a job would cross the line, not after. Kokoro is exempt from the meter, so drafting and regenerating never stop. Your library and exports stay exactly where they are.
Can I use the free audio commercially?
Yes. Ownership here does not depend on the price you pay: audio generated on the free plan is yours to publish and monetize, and the paid cards add the printed commercial rights, no watermark line for client-grade work. The plain-language license lives on the ownership page.
Which engines does the free plan include?
All five engines are in the studio from day one. Kokoro runs without a meter, and your premium minutes let you hear the expressive engines, Gemini Flash included, perform your own script before any upgrade decision.
Keep exploring

Put your own script through it.

An email gets you in, the receipt above is the whole deal, and the first file you export is yours to keep.

Start freeExplore Text to Speech