Private enterprise data
structured for AI
Access hard-to-source, de-identified enterprise datasets optimized for AI model development.
We preserve the relationships between entities, actions and outcomes to produce structured datasets that are ideal for building training tasks and evaluations.
Data sources
Enterprise operations
Business records, internal documents, communications and operational histories from traditional small and mid-sized enterprises.
Software and technical archives
Source code, technical documentation and historical development records from startups, including ones that have wound down.
Specialized data collections
Domain-specific datasets, including private rare-language codebases, sourced through enterprise relationships and authorized asset acquisitions.
Streamlined licensing
- Zamorak licenses data in its own name, under authority from our data providers, so there is no separate negotiation with individual data owners.
- Each dataset undergoes rights, provenance and privacy diligence, with documented licensing terms to simplify procurement and reduce acquisition risk.
Datasheet
Template- Source
- the systems the records come from
- Rights
- owned by Zamorak, or licensed to Zamorak by its owner
- Period
- the licensed window
- Records
- counted at delivery
- Modalities
- text, attachments, structured exports
- De-identification
- method named in the agreement
- Expert determination
- when required
- License
- nonexclusive; term and field of use
- Format
- machine-readable, documented
Absent: removed before delivery
- Names
- absent
- Email addresses
- absent
- Phone numbers
- absent
- Free-text identifiers
- absent
Built around your requirements
We are responsive to your specifications. Work with our team to identify available sources, evaluate representative samples, and agree on delivery formats and licensing terms.
Or write to data@zamorak.ai