Product updates, new features, improvements, and fixes as we ship them. Subscribe via RSS to stay updated.
This changelog reflects the current mainline release. Availability in your environment depends on your organization's deployment cadence.
Administrators can now assign compliance grants to bundles directly from the admin panel without waiting on automated resolution. The underlying data model handles conflict resolution so grants stay consistent across the platform. A dedicated bundle authoring interface lets admins compose and publish grant assignments from a single place.
Administrators can now move a governance workflow into any state directly from the Governance panel and apply a backdated timestamp when doing so. This makes it straightforward to correct workflow state after the fact or record approvals that happened outside the system. The backdated date travels with the state change so audit records stay accurate.
Human judging sessions now support entry sampling and filtering, so reviewers work through a representative slice rather than every item in a run. Inter-rater reliability (IRR) scoring using Cohen's kappa shows how consistently reviewers agree, making calibration gaps visible before they affect results. Results and disagreements export to CSV for offline analysis, audit records, and integration with existing reporting workflows.
Claude Haiku 4.5 from Anthropic is now available through the Swept model gateway. The model is accessible via the same unified, OpenAI-compatible endpoint as other supported models and includes routing support for Claude Code integrations.
The two monthly AI spend tiers have been renamed to better reflect how teams use them: Advanced is now Power User (Max) and Max is now Developer (Max). Monthly spend limits have also increased, with Power User moving from $15 to $20 and Developer moving from $25 to $40.
The admin users view now shows live AI token spend for every user alongside their assigned budget tier. Administrators can reassign a tier without leaving the page, and each change is recorded in the audit timeline. Combined with the existing per-user budget limits, this closes the loop between who is spending, how much, and what ceiling applies.
The AI workspace now extracts document content using Apache Tika, enabling richer file types such as Word documents and presentations to reach the model with full fidelity. PostScript files are automatically converted to PDF before extraction, removing a common friction point for teams that work with print-ready output. Documents are injected in full rather than retrieved in fragments, so the model has complete context when answering questions about uploaded files.
User profiles in the admin panel now display the OIDC directory groups that grant access to platform features, alongside the timestamp of each user's most recent sign-in. The data comes directly from your identity provider, so group membership reflects the source of truth without manual reconciliation. Teams running regular access reviews can now pull this context without leaving the admin UI.
Playgrounds on Demand lets you provision a short-lived, fully-isolated copy of your environment whenever you need to run a demo, test a configuration change, or try an experiment. Each playground deploys from the current production image and tears down automatically when the session ends. There is no standing infrastructure to maintain and no risk of side effects on live data.
The AI Application detail page now has a dedicated Changelog tab that shows what changed in each release. When a deployment is waiting for approval, reviewers also see a structured summary of the changes it contains so they can assess impact before the release goes live. Both additions support change-control policies and make it straightforward to maintain a clear audit trail without leaving the platform.