Turn your company's engineering history into an asset.
Ofton buys anonymised internal data, such as code, tickets, documents and design files, to build AI training datasets and benchmarks. Describe what you have below. If it's a fit, we'll schedule a call.
- Descriptions only, no files
- NDA before any sample
- Only anonymised output is used
- At most 10% AI-generated
- About 10 minutes
Human-made data only: no more than 10% of the content can be AI-generated (you'll estimate it in section 04). Descriptions only: don't paste any actual data, such as code, messages, names or credentials.
Common questions.
What exactly are we selling?
An anonymised dataset built from your data, defined precisely in the sale agreement. Your original systems and files stay with you.
What happens to personal data?
It is removed or replaced before any data is used. Names, emails, credentials and other identifiers are stripped and the output is checked before it enters a dataset. Data about minors, health data and data from sensitive sectors is excluded.
Can we stop at any stage?
Yes. Nothing is binding until you sign a sale agreement. You can stop after the form, the call or the sample review.
Can the data include AI-generated content?
Only a little: no more than 10% of the content. We buy data to show how people really work, so it must be mostly human-made. If your team adopted AI tools recently, the work made before that date is the most useful. Tell us your estimate in section 04, and we'll check it on the sample.
How is the price set?
By the volume, history and quality of the data, and how much of it is usable after anonymisation. Data with strong quality signals is priced higher.