Foní
Log in Sign up

Own your voices. Stop renting them.

The human data to train your own TTS: licensed native voices, expert accent reviewers and audio labelers. Pay per unit, no per-character fees.

Work addresses only. We don't accept personal email like Gmail or Outlook.

Listen: three voices, three accents 0:00 / –:––

Voices are listed by first name, initial and catalog number. No photos.

Everything a voice model needs from humans.

Voices

Native voice talent, sourced to your brief, recorded and licensed to you in writing.

2-minute sampleFree
15-minute recording$100
90-minute session$600
License$6,000

Expert reviewers

Native experts score your audio for accent, delivery and quality. Up to three reviewers per sample.

Per sample, per reviewer$12

Labelers

Native labelers transcribe and tag the audio you already have, so it is ready to train on.

Per labeled audio hour[PRICE]

From public audio to licensed voices, without a weekly call.

Lane 1 · Base model · teach your model a language from audio you collected

  1. 1Upload clipsEvery file gets an automatic audio-quality check before anyone is paid to touch it.
  2. 2LabelNative labelers transcribe and tag each clip.
  3. 3ReviewExpert reviewers validate language, accent and quality.

Lane 2 · Product voices · the voices your agents ship with

  1. 1FreeRequestLanguage, accent, gender and how many voices you need.
  2. 2FreeReview samplesListen to 2-minute samples, already scored by experts. Advance, ask for another round, or reject.
  3. 3PaidRecordOrder a 15-minute recording, then a 90-minute session, one click each.
  4. 4PaidLicenseSigned by the talent. The certificate and the audio stay in your account.

Native experts for every accent, already vetted.

Is that really a Monterrey accent? Ask someone from Monterrey.

Entry testEvery expert scores reference samples before they get paid work.
Three reviewers per audioYou get the median score and a flag where they disagree.
One rubricFive criteria scored 1 to 5: conversational delivery, authenticity, contextual intention and emotion, pacing and flow, language and accent fit.
The bench todayReviewers
Spanish · Colombian[N]
Spanish · Mexican[N]
Portuguese · Brazilian[N]
Arabic[N]
French[N]
A typical reviewerNative of Barranquilla, translator, 4 years reviewing

Native labelers for the audio you already have.

What they labelTranscription and tags on every clip. [LIST THE LABEL TYPES YOU OFFER]
LanguagesEach clip goes to a native speaker of its language. [LANGUAGES WITH LABELERS TODAY]
Quality controlExpert reviewers validate language, accent and quality. You accept or return each batch.

Pay per unit.

Sourcing and 2-minute samples are free. You pay for recordings, licenses, reviews and labeled hours.

ItemUnitPrice
Sourcing and 2-minute samplesPer request, both roundsFree
15-minute recordingPer voice$100
90-minute sessionPer voice$600
LicensePer voice$6,000
Expert reviewPer sample, per reviewer$12
Labeled audioPer audio hour[PRICE]

Pay by saved card. No subscriptions, no minimums, no per-character fees.

18 languages, 24 named accents, 600+ reviewed voices.

Need one that is not here? Request it.

Every voice is recorded with consent and licensed in writing.

Signed talent contractsEvery voice talent signs before their audio reaches you.
A clear license for youScope, date and voice on one certificate you can download any time.
No public linksAudio is only available inside your account, never on an open URL.

Buying at volume? Pay by monthly invoice.

Enterprise accounts get contract pricing and dedicated reviewer and labeler teams.

Talk to us

Questions

What is free?

Creating a request, sourcing talent, and every 2-minute sample, including a second round. You start paying when you order a 15-minute recording.

How long does it take?

First samples in [X] days. A 15-minute recording in [X] days, a 90-minute session in [X] days. Every order shows its promised date.

What does the license cover?

[LICENSE SCOPE: uses, term, territory, exclusivity]

Which audio formats do you deliver and accept?

[DELIVERY FORMAT, SAMPLE RATE, BIT DEPTH] for recordings. [ACCEPTED UPLOAD FORMATS] for clips you send to review or labeling.

Who are the experts?

Native speakers of the accent they review, such as translators and linguists, who passed an entry test on reference samples. You see their origin, background and score, never a name or photo.

Who can sign up?

Teams building voice products, with a work email. Personal addresses aren't accepted. If your company already has an account, you'll be asked to join it.

Your first samples are free.