What we collect and license
Speech in any language
Prompted and conversational speech, dialects and low-resource languages, with transcripts and speaker metadata.
Text
Native-written text, translations, instructions and domain writing in the languages your models are weak in.
Images and video
Photos and video of real places, objects and activities, captured to your brief in the countries you need.
First-person and robot data
Egocentric video of everyday and work tasks, and demonstration data for physical-AI and robotics teams.
Human feedback and labels
Native-speaker review, rating, annotation and private evaluation sets your models have never seen.
Licensed datasets
Archives and datasets from publishers, broadcasters and other data owners, cleared for AI training.
Not on this list? We welcome data of every kind: business, financial, retail, geospatial, scientific, sensor, legal, health (de-identified) and more. If you own data, we can license and sell it for you. If you need data, we will find it, collect it or license it.
How we work
- You tell us the spec. Languages, countries, data type, volume, format, deadline, and a shared or exclusive licence.
- We recruit contributors. Native speakers and local contributors, found through universities, community groups and partners in each region.
- Every contributor consents and is paid. They see what the data is for and what they earn before they agree. We keep a versioned consent record for each person.
- We check and deliver. Automatic and human quality checks, then delivery with a manifest carrying the consent and origin record for every file.
Languages we cover
We track how much open speech data exists for 445 languages, and collect new data in any of them. See the gaps for your language.
Partner with us
We partner with anyone: data owners, companies, governments, universities, researchers, communities, resellers, marketplaces and individuals. If you have data, need data or can help collect it, talk to us.
Data owners
Own recordings, text, images or video? We license them to AI teams with clear terms and share the revenue with you.
License your data →Communities and universities
Language associations, student groups, NGOs and schools: help your members earn from paid recording and data work.
Become a partner →Contributors
Speak a language or have a skill AI needs? Join the waitlist and we'll email you paid projects that fit.
Join as a contributor →Anyone else
Buyers, resellers, marketplaces, startups, labs and agencies: if there is a way to work together, we want to hear it.
Email us →Questions
What data does Norynta provide?
Consented AI training data: speech, text, images, video, first-person and robot-demonstration data, and human feedback. We collect it to your specification in any language or country, and we license existing datasets from data owners.
Which languages do you support?
Any language with native speakers we can reach, including dialects and low-resource languages. Our public index tracks open speech data for 445 languages so you can see where the gaps are.
How is consent handled?
Every contributor sees what the data is for and what they will be paid before agreeing in the app. Each file is delivered with a manifest that links it to a versioned consent and origin record.
Is the licence shared or exclusive?
Both. Datasets are non-exclusive by default; custom collections can be exclusive to you.
How can I earn as a contributor?
Join the waitlist at collect.norynta.com/join. We email you paid projects in your language, and you record or write on your phone. You are paid once your work is accepted.
What kinds of data do you buy and sell?
Any kind, as long as it was collected lawfully and the owner has the right to share it: AI training data, business and market data, geospatial, scientific, sensor, financial, legal, de-identified health data and more. We license data from owners and sell or license it to buyers, with the rights and consent record kept for every file.
How is pricing set?
Openly. For custom collection you see what each contributor is paid, plus one published Norynta fee of 30% that covers recruiting, quality checks, consent records and delivery. Ready datasets are licensed at a fixed quoted price. Work starts after the first payment or a signed contract.
Start a project or ask a question
Tell us what you need, or what you'd like to offer. You'll get a personal reply within one business day.
Or email data@norynta.com.