Licensing guide
How to license GitLab data to AI companies
Quick answer
GitLab holds merge requests, reviews, pipelines and issues. AI companies pay for this kind of real work history because it shows how problems are actually solved. Through SourceX, you export a copy, personal details are removed and checked, you approve the final package, and only then is it licensed under terms you sign.
Why AI companies want GitLab records#
- These records show how real work gets done, which is what AI companies cannot create synthetically
- Buyers use them to train and evaluate AI systems on real-world tasks
- Several years of consistent history carry the most value
How licensing works#
- Take the fit check to see whether your GitLab history matches current buyer demand
- Export a copy of your records — our step-by-step GitLab export guide walks through it
- SourceX prepares the package: personal details removed, samples reviewed by a human
- You review and approve the final package before anything is shared
- Terms are agreed in writing — approved uses, exclusivity and payment — before delivery
- Large datasets stay in your own storage or ship on encrypted drives; SourceX never needs access to your live systems
Start with the export
The first practical step is a complete export of your GitLab history. Our guide follows GitLab's own documentation.
Read the GitLab export guideSources
Public sources, checked 2026-10-09. SourceX is not affiliated with the organizations named.
- GitLab: Project import and export docs.gitlab.com
See if your company qualifies
A short company assessment. No data uploads are needed.