Release Notes 02.04.2026 - v3.27.0
📝 This release focuses on improving system stability, enhancing performance, and expanding key platform capabilities across backend, frontend, and AI-related features. Several improvements were made to data handling, monitoring, and user experience, along with important bug fixes to ensure smoother workflows.
📅 02.04.2026
🔖 Version 3.27.0
✨ Features
- Improve height adjustment of the process view
The process view now supports better height adjustments, making it more flexible and improving usability across different screen sizes. - Show statistics in properties
Added visibility of key statistics directly within properties, allowing users to quickly access important insights without navigating away. - Introduce warning if user wants to discard chat edit
Users now receive a confirmation warning before discarding chat edits, reducing the risk of accidental data loss. - Save agent meta information in DB
Agent metadata is now persisted in the database, enabling better tracking, management, and future extensibility. - Show token consumption and used context
Users can now view token usage and context consumption, improving transparency and optimization of AI interactions. - Add entity JSON generator as external HTML tool
Added a new tool to generate entity JSON externally, enabling easier integrations and data handling. - Adjust prompts to refine alternative suggestions handling
Improved prompt behavior to provide more relevant and controlled alternative suggestions. - Dynamic agent registry
Introduced a dynamic registry for agents, allowing more flexible configuration and runtime adaptability. - Enable web search for OpenAI integration
Introduced web search capability to enhance AI responses with more up-to-date and relevant information. - Data Optimisation
Introduced a data optimisation approach to improve how data is structured and retrieved. This reduces response sizes, improves loading times, and simplifies request complexity while establishing a stronger foundation for faster and more scalable feature development.
🐛 Bug Fixes
- Cache handling improved when changing datasource configuration
Fixed issues with cache not updating correctly after datasource changes, ensuring consistent data behavior. - Improve AI center tokenize endpoint configuration handling
Fixed issues related to hardcoded embedding URLs, making the system more flexible and reliable. - Logout button for Minerva Chat improved
Addressed issues with logout functionality to ensure a smoother and more predictable user experience. - Entity graph config zoom behaviour improved after edits
Fixed inconsistencies where zoom levels were not preserved correctly after changes. - Token statistics accuracy improved
Fixed incorrect token statistics to provide more reliable usage insights. - Violation frequency slider behaviour improved
Addressed unexpected behavior in the slider to ensure smoother and more predictable interactions. - Entity-graph relationships persistence improved without mapping issues
Resolved problems with saving and testing relationships, ensuring consistent data persistence.