ZBrain Release Notes
Last updated
Last updated
Release Date: May 23, 2025
The ZBrain 2.3.0 release delivers major enhancements focused on observability, user control, and platform transparency. Highlights include agent activity monitoring, user-based filtering in Flows & Pieces, a new "About" page for integration clarity, expanded embedding support in RAG, and improved sorting, search, and metric integration in the evaluation framework.
ZBrain About Page: Added an “About” section to enhance user understanding of ZBrain integration, providing context and key capabilities.
Tip: Navigate to Settings → About ZBrain to explore key platform details, including version info, legal resources, and privacy policies, your one-stop overview of ZBrain’s foundation and compliance.
Flows & Pieces
Agent Monitoring: Enables real-time visibility into agent activity across workflows.
User-based Filtering: Allows filtering of flows, connections, and runs by individual users for more granular monitoring.
Evaluation Framework
Search & Sort Enhancements: Added sorting and search capabilities for easier data navigation.
New Metrics Integration: Enabled integration with the prompt manager, supporting custom evaluation metrics.
RAG (Retrieval-Augmented Generation)
Expanded Embedding Models: Expanded embedding model support for improved ingestion flexibility and retrieval precision.
Knowledge Graph Enhancements: Refined knowledge graph creation logic for cleaner entity-relationship mapping.
UI Upgrades: Delivered a more intuitive and interactive UI for managing RAG-based knowledge assets.
Agentic Framework
Agent Management Enhancements: Enabled tracking and management of agents within active pipelines.
Response Logging: Response Logging: Introduced logs to track agent outputs and reasoning steps.
Mastra LangGraph Integration: Integrated with Mastra LangGraph for orchestrating multi-agent workflows using graph-based logic.
CoI (Center of Intelligence)
Notification System: Introduced a bell icon in the interface to display real-time system alerts and notifications, enhancing user awareness and responsiveness within the platform.
Production Deployment: Successfully released the COI platform to the production environment, marking a key milestone toward live platform operations and user access.