Haitian Creole speech data.
Recorded by Haitian Creole speakers in Haiti and abroad. Transcribed and checked by people. Built for AI companies that want their products to work in Haitian Creole.
Our first collection starts soon. We’re already taking requests for evaluation sets and custom collections.
M pral nan mache a talè.
“I’m going to the market soon.”
One sentence. One speaker. Checked by two reviewers.
What we provide.
More than twelve million people speak Haitian Creole. Most voice AI still can’t understand it, because the data that exists is old, narrow or unconsented. We collect the Creole people speak today.
Evaluation sets
A few carefully checked hours, balanced across regions, to measure how well your model handles Haitian Creole.
Training data
Natural speech and read speech, with human transcripts in the official Haitian Creole spelling.
Custom collections
Your topics, regions or scenarios, such as medical conversations or customer calls.
Voice data
For text-to-speech, only from speakers who said yes to voice generation.
More creole languages.
We’re starting with Haitian Creole. Next, we plan pilots in other creole languages that AI still leaves behind, from the Caribbean and the Indian Ocean to Cabo Verde and West Africa.
- Antillean Creole (Guadeloupe, Martinique, Dominica, Saint Lucia)
- Guianese Creole
- Réunion Creole
- Mauritian Creole
- Seychellois Creole
- Cape Verdean Creole (Kriolu)
- Guinea-Bissau Creole (Kriol)
- Papiamento
- Jamaican Patois
- Krio (Sierra Leone)
- Belizean Kriol
Speak one of them? Join the listNeed data in one of them? Write to us
How it works.
Every recording passes four steps before it counts. Nobody checks their own work.
Record
Uncompressed audio from the speaker’s phone, with noise filters off.
Check
Automatic checks for length, silence, clipping and duplicates.
Transcribe
A reviewer checks the audio is clear and on topic, then writes exactly what was said.
Validate
A second reviewer compares the transcript with the audio.
We re-check a random 5% of finished recordings ourselves.
Consent and fair pay.
The people behind the data come first. That’s also what makes it safe to use.
- Every contributor is 18 or older and agrees in their own language before recording.
- Voice generation is a separate yes or no.
- Contributors can stop at any time and have their recordings deleted.
- No names or contact details ever reach buyers. Speakers get a random code.
- Paid for every approved minute and every review task.
- Paid weekly by MonCash or NatCash in Haiti. Payment abroad (Wise or PayPal) coming soon.
- Reviewers earn the same whether they approve or reject.
- People who leave are still paid for everything they earned.
Do you speak a creole language?
Piplis pays Haitian Creole speakers to record their voice with their phone. We also need people who read and write Haitian Creole well to review recordings.
In Haiti, payment by MonCash or NatCash. Payment for people abroad is coming soon.
Speak another creole? Join the list too: we’ll contact you when a pilot opens for your language.
Sign-ups open soon. Leave your details and we’ll email you when they do.
Let’s talk.
Datasets, evaluation sets, custom collections or partnerships. Need data in another creole language? Tell us which one.
[email protected]Jehud Neri Sabbat, founder