Data Hub
Reference guide for the Data Hub — audit trail, versioning, and AI Act evidence.
Overview
The Data Hub is your workspace’s audit trail and version control system. It tracks every change made to datasets, documents, prompts, models, runs, and ground truths — providing full traceability for compliance and auditing.
Folder Structure
The Data Hub organizes history into six expandable sections:
| Folder | What’s Tracked |
|---|---|
| Datasets | Creation, updates, deletion |
| Documents | Uploads, renames, moves, deletions |
| Prompts | Creation, updates, archival |
| Models | Additions, updates, enable/disable |
| Runs | Benchmark creation, completion, failures |
| Ground Truths | Label creation, updates, deletion |
Each folder lets you search by name or version and lists entries with:
- Version number (v1, v2, etc.)
- Timestamp
- Action label — Create, Update, Delete, System, or Snapshot
- Color coding by action type
Timeline
The right section shows a chronological timeline of recent changes across all entity types. Each entry shows the entity type, what action was taken, who performed it (user or system), and when.
AI Act Evidence Pack
The Data Hub includes an AI Act Evidence Pack feature for EU AI Act audit evidence:
This feature is currently being built, and will release shortly. In the meantime, all of your data is captured and will be retroactively cast into this feature, at no extra cost.
This feature helps you build the documentation trail required for regulatory compliance.