Vocalix speech platform · text to speech
Speech that doesn’t drift under you.
78 voices across 10 languages — deepest in English and Hindi, including the easy mix of the two that most voices stumble over. The voice you launch with is the voice you keep. Run it on our servers, or on your own so nothing you record ever leaves your building.
No card needed · free to try · nothing charged while you build
01 — Try it
Hear it in the browser, before you sign up.
Pick a Hindi sentence, pick a voice, press play. The audio is generated when you press it — no account, no card, nothing to install. More about the Hindi voices.
02 — Proof
Promises we can actually keep.
Every demo clip sounds good. These are the things that decide whether it still sounds good a year in, and whether finance ever gets a surprise.
Consistency
Say it twice. Sounds the same.
The same sentence in the same voice gives you the same audio, this week and next year. Nothing shifts quietly underneath a product you have already shipped, and your own tests never wobble.
Two languages at once
Mixed sentences, handled
आपका parcel deliver हो गया
Switching between Hindi and English mid-sentence is the normal case here, not a special one — in both directions.
Stability
Your voice stays put
Choose a voice once and it is locked to your account. A new release on our side can never change how your product sounds on a Friday afternoon.
Billing
No surprise invoices
You pay from a balance you top up, and we tell you before a job runs if it would take you past it. There is no bill at the end of the month you did not expect.
Repeats
Say it again for less
Phrases you use over and over — greetings, menus, confirmations — cost a fifth of the normal rate after the first time.
Security
Your account, your control
Access is shown to you once and stored encrypted, never in a form anyone here can read back. Turn it off and it stops working immediately. Your audio and text are never used to train anything.
03 — Where it runs
Same service, three places to run it.
Most speech companies sell you one: their building, their network, their copy of your recordings. You get the same voices and the same price whichever you choose. What changes is where the audio is handled — usually the only question a security review really cares about.
We run it
Our servers
Nothing to install and nothing to maintain. Send us the words, get the audio back. This is what most people use.
Direct connection
A machine near your users
For high volume and the fastest possible response, your app connects straight to a machine near your users. We handle the billing quietly in the background.
You run it
Your own building
Everything runs inside your building, on your machines. Your recordings never reach us — only a running total, so we can bill you.
On your own hardware
Nothing you record ever leaves
For teams who are not allowed to send recordings outside the building — hospitals, banks, government work — everything runs on your machines, inside your network. The only thing that ever reaches us is a running total for billing. We never receive the audio, and there is no way in from our side.
04 — Voices
78 voices, and room to be picky.
The library leans heavily on English and Hindi rather than spreading itself thin, so here is the honest count for each language instead of a row of flags. Every voice costs the same and behaves the same.
| Language | Voices | Accents |
|---|---|---|
| English | 44 | American · Indian · British · Australian |
| Hindi | 18 | Indian · Gujarati · Bengali · Marathi |
| Spanish | 4 | Latin American · Peruvian |
| Tamil | 3 | Indian |
| Korean | 2 | Seoul |
| Arabic | 2 | Saudi · Jordanian |
| Portuguese | 2 | Brazilian |
| Italian | 1 | Tuscan |
| French | 1 | African |
| Chinese | 1 | Beijing Mandarin |
05 — Price
Pay for what you use. No seats.
No monthly minimum, no per-user fee, and credit that never expires. The price is the same whether it runs on our machines or yours. Start with $5 on us.
| What you get | Price |
|---|---|
| Text to speech — per minute of audio | |
Text to speech Every voice in the library, at the same price. | $0.0189 /min |
| Speech to text — per minute of audio | |
Not open yet Built, and free to try today, but not yet switched on for paid use — so we do not sell it yet. | — |
Testing while you build Unlimited, and never charged | Free |
06 — For your team
Your engineers will be done in an afternoon.
It is a handful of lines in whatever you already write, and if your team already builds with OpenAI’s tools, switching over is a one-line change. Everything they need — setup, examples, a free way to test the whole thing before spending anything — is in the documentation.
Ship a voice your users recognise.
Five minutes to your first sentence. Credit on signup, no card, and a free way to keep testing while you build.