Changelog
A running list of what we have shipped. We post the changes that affect how you work, not every internal tweak.
One click audit log verification
Verify the whole hash chain from the dashboard. If any past entry was changed, the check points to the exact row where the chain breaks, so you can trust the record or find the problem fast.
MCP server is generally available
Connect agents that speak the Model Context Protocol without writing custom code. Point the server at your API key and actions start flowing into the inventory and the audit log.
Five evaluation suites
Bias, hallucination, prompt injection, safety, and policy checks now run as separate suites. Each has its own score and its own pass threshold, so a release can be gated on the tests that matter for that system.
Risk classification with source citations
Every risk tier now links to the exact EU AI Act clause behind it. A reviewer can trace each decision by hand, and the reasoning is stored next to the system it describes.
Document generator
Generate risk assessments, Annex IV technical files, model cards, and audit reports straight from your live data. When a system changes, regenerate the document in a click rather than editing it by hand.