Trace software to its source
Models inherit the rights of the data behind them.
A fine-tuned model is an asset whose defensibility depends on the provenance of its training data and the terms of its base model. Recording those relationships now is considerably cheaper than reconstructing them during a raise.
AI Models & Datasets
- Model records & lineage
- Dataset provenance
- Usage rights & restrictions
- Model-to-product relationships
How it helps
What you get with AI Models & Datasets
Model records with lineage
Base model, fine-tuning runs and evaluation sets recorded as related objects rather than as a filename.
Dataset provenance
Where a dataset came from, under which licence or consent, and what it may be used for.
Usage rights travel
Models connect to the products that embed them, so a restriction on a dataset is visible from the product it constrains.
Software Provenance
The rest of Software Provenance
Connect repositories, contributors, agreements, components, releases and SBOM data to legal ownership.
Repositories
Inventory repositories, organizations, releases and technical custody.
Learn moreContributors & Agreements
Connect contributor identities and contribution periods to the agreements that cover their work.
Learn moreSoftware Products & Components
Model proprietary products, shared core components and brand-specific derivatives separately.
Learn moreSBOM & Third-Party Components
Import open-source and third-party component data from established standards.
Learn moreBuild the ownership record before somebody asks for it.
Register what exists, connect the relationships, attach the evidence. The answer you need under deadline is the one you maintained beforehand.
Evidence-backed relationships Multi-entity by design Built for transfer and diligence