01
The product
The Data Hub is where it all starts: native speakers record words and phrases, transcribe audio, and label images, then moderators validate every contribution before it trains a model. Every accepted contribution is paid, and a leaderboard follows the community's activity. Companies and institutions can then access this structured data under license.
02
The product story
Three ways to contribute
The Data Hub breaks collection into short tasks you can do from a phone. You pick the one that matches what you can do, in the language or languages you speak.
- Record — you read a word or a phrase, the platform captures your voice.
- Transcribe — you listen to existing audio and write down what you hear.
- Label images — you describe or classify an image in your own language.
Nothing enters a dataset unreviewed
Every contribution goes past a native-speaker moderator before it is kept. The platform splits the work across five roles — contributor, moderator, collector, accountant and administrator — so the person who produces a piece of data is never the one who validates it.
Data that no native speaker has reviewed is not training data.Data Hub design principle
Accepted work is paid
Every validated contribution is paid at the current rate, and a dedicated role follows the payout. A leaderboard makes the community's activity visible, and performance tracking shows what the collected data actually changes for the model.
03
Contribute
What a contributor does, step by step.
Create a contributor account
Sign up on the Data Hub with the language or languages you speak.
Pick a task
Record a word or a phrase, transcribe existing audio, or label images.
A moderator reviews it
A native moderator checks and validates your contribution before it enters the dataset.
You get paid
Every accepted contribution is paid at the platform's current rate, and the leaderboard follows your activity.
04
Become a client
How data access works, from request to delivery.
Specify your needs
Languages, domains (health, legal, conversational), volume, and data format.
We scope a license agreement
Access to existing datasets or a dedicated collection, under a clear contract.
Delivery and support
Structured data export, with support to integrate it into your pipelines.