Release Notes 23.10.2026 - v5.3.0
📝 This release introduces a completely reworked tab experience: analyzer tabs and dashboards are now organised in an explorer sidebar with folders, favourites and search. Imports become far more transparent with a detailed, step-by-step report for every perspective and entity graph, and perspectives with many cases or complex causal models load considerably faster. Minerva shows what each AI model costs and records token usage reliably, and every object now shows who owns it. Dark mode follows your system setting, a new, more detailed tool documentation is available at docs.noreja.com, and several fixes round off the release.
📅 23.10.2026
🔖 Version 5.3.0
✨ Features
- Reworked tab management
Analyzer tabs and dashboards now live in an explorer sidebar where you can sort them into folders, mark favourites, nest tabs under each other and find them quickly with search and filters. A slimmer tab bar, drag and drop, context menus and keyboard navigation make working with many tabs easier. Changes to filters and settings are saved automatically, tabs you haven't opened for 14 days are suggested for clean-up, and the last opened tab is restored when you come back. - Detailed import report
Clicking a perspective or entity graph in the builder opens a side panel with a step-by-step report of its latest import, including the duration of each step and, if something went wrong, where it failed together with the full error message. The table itself stays compact, showing the outcome and progress of the current import at a glance. Previous imports are listed as well, and entity graphs also show how many activities and relationships were imported. - Builder dialogs as side panels
Creating, copying, editing and importing perspectives and entity graphs now happens in side panels instead of pop-up dialogs, so the overview stays visible while you work. - Faster perspective loading
Perspectives with very many cases or long, branching causal models now load in considerably less time, and load times are more predictable. Violations are calculated more efficiently, while the process graph shows the same results as before. Causal models that cannot be processed now stop the load with a clear message naming the affected case. - Object owners at a glance
Every object in access management, as well as agents and prompts, now shows who owns or shared it. If you need more permissions on an object, you can see straight away whom to ask. - Price indication for AI models
When choosing a model for an agent or a chat, a symbol now indicates how expensive it is, and the exact prices are shown on hover. Selecting one of the most expensive models asks for confirmation first. - Accurate token usage and costs
Minerva now records the tokens and costs of every request separately, so the usage history stays complete even when conversations or messages are deleted or regenerated. Prompt caching lowers the cost of longer conversations, and administrators can retrieve the usage statistics directly. - Richer context search
Context search results now include the document summary, the chapter summary, the folder name and the matching text passages, making it easier to judge the relevance of a result without opening the document. - Automatic dark and light mode
The colour mode now follows your operating system or browser setting automatically, and your chosen mode is kept after logging out. - New tool documentation
A new, more detailed and up-to-date documentation of the tool is now available at docs.noreja.com. It describes the features of the current version and is kept up to date with each release. - Workbench management in the settings
The overview of workbench instances has moved from the builder menu to the settings, next to the other environment configuration.
🐛 Bug Fixes
- Minerva respects access rights for perspectives
When Minerva lists perspectives and tabs, it now only includes those you have access to. Shared analyzer tabs are analysed with their own filters, just as in the analyzer. - Correct relationship counts in Minerva
Minerva now recognises when violations are aggregated, so the relationship counts it reports match those shown in the process graph. - Unread chat conversations are remembered
The unread state of chat conversations is now stored on the server and persists across logouts and browser restarts. - Reliable resizing of the workspace and chat panels
The workspace and chat panels can be resized smoothly again without snapping back to a smaller size. - New chat button works from any conversation
Starting a new chat now works directly from within an open conversation. - Dashboard widgets follow a colour reset
After resetting a custom theme colour, dashboard widgets now return to the default colours as well.
🔮 Coming Soon
- Customer Branding for Minerva-Generated Charts
Charts generated by Minerva will follow your company's branding, ensuring a consistent visual identity across analyses and reports. - Improved Transition Duration Widget
The Transition Duration Widget will receive improvements to make it easier to explore and understand the time spent between activities, providing clearer insights into process performance. - Support for More Data Types
Minerva will support a broader range of data types, making it possible to work with more kinds of data across analyses and connected data sources. - Update Confidential Context
Confidential context will be updateable, allowing existing protected company knowledge to be kept current without having to recreate it. - Further UI/UX Improvements
More improvements are on the way to make the platform easier and more intuitive to use. Expect smoother interactions, more consistent designs, and refined workflows for a better overall user experience.