1. Agents Kit 2.2 – Human-in-the-Loop Controls
Mendix 11.14 enhances agentic application development with Agents Kit 2.2, introducing Human-In-The-Loop controls in the Agent Editor. Developers can now configure how end users interact with agent tools and knowledge bases by choosing whether specific tools are visible, hidden, or require approval before invocation. Teams can also customize the tool names and descriptions displayed to users.
These capabilities provide organizations with greater control over agent behavior while keeping human oversight at the center of AI-driven workflows. By defining appropriate visibility and approval requirements, development teams can create agentic applications that provide users with transparency and control over AI actions, helping organizations implement AI capabilities in a more governed and user-aware manner.
Key Takeaway: Human-In-The-Loop controls provide greater visibility, governance, and user control over AI agents.
2. Real-Time Response Streaming and New Claude Models
Mendix 11.14 introduces response streaming for agents, allowing AI-generated responses to appear word by word in real time instead of making users wait for the complete response. The capability works with the Conversational UI module and is supported across Mistral, Google Gemini, OpenAI, and Mendix Cloud GenAI connectors, with Amazon Bedrock support planned for a future release.
The release also adds Anthropic Claude Sonnet 5 and Claude Opus 5 to Mendix Cloud GenAI Resource Packs. Developers can access these newer Claude models directly through the GenAI Text Generation resource without reworking their existing application setup, providing access to the latest Claude generation while maintaining the same GenAI Unit exchange rates as other models in their respective families.
Key Takeaway: Streaming responses and newer Claude models create faster and more capable AI application experiences.
3. Maia Units – Usage Insights and Budget Management
Mendix 11.14 introduces enhanced visibility into Maia usage through Maia Units, giving organizations a real-time view of how AI development resources are being consumed. Through the Control Center Maia page, administrators can view total units, consumption levels, capacity allocation, and usage at the individual-user level.
This visibility helps organizations manage Maia resources more effectively and keep usage within defined limits. Company administrators can review configured limits and identify how resources are being used across the organization, while teams can provision additional resources using cloud tokens when they need to scale their Maia usage and support broader AI-assisted development initiatives.
Key Takeaway: Maia Units provide organizations with greater visibility and control over AI development usage.
4. Improved Developer Navigation, Search and Branch Management
Mendix 11.14 improves developer productivity with enhancements to the History pane, search experience, and branch management. Developers can now inspect the history of other branches without downloading the complete branch and can create a new branch directly from a selected commit, making it easier to isolate and work on specific project versions.
The redesigned Find Results panel also introduces a History Bar that automatically maintains previous searches. Developers can quickly switch between current and earlier search results, revisit previous investigations, and remove searches they no longer need. Together, these improvements simplify application navigation, make code and model investigation more efficient, and support smoother collaboration across development teams.
Key Takeaway: Enhanced history, search, and branch capabilities make Mendix development more efficient and collaborative.
5. Java 25 Support and Improved Offline Data Management
Mendix 11.14 adds support for Java 25, providing developers with access to a newer stable Java runtime and associated performance improvements. Internal testing cited in the release showed Java 25 consistently outperforming Java 21 across a range of benchmarks, while Java’s backward compatibility makes the transition relatively straightforward for most applications. New projects use Java 25 by default, while existing applications can be upgraded through project settings.
The release also improves offline-first application management with enhanced Clear from Device functionality. Developers can now remove individual objects or lists of objects from local device storage instead of clearing every object belonging to an entity. This provides more granular control over offline data, helping applications manage device storage more efficiently while ensuring that removing local data does not delete corresponding objects from the server.
Key Takeaway: Java 25 and granular offline data controls improve application performance and mobile development flexibility.
6. Enterprise Certificate Management and Database Monitoring
Mendix 11.14 expands enterprise capabilities with Certificate Management APIs that allow organizations to list, create, update, replace, and delete centrally managed SSL/TLS certificates. These APIs can integrate certificate management with existing enterprise tools and help automate the certificate lifecycle, reducing the risk of operational disruption when certificates need to be renewed or replaced.
The release also introduces three advanced database metrics for Mendix Cloud: Average Database Bloat Ratio, Maximum Table Bloat Ratio, and Largest Table Size. These metrics provide greater visibility into database health and can help teams identify issues such as excessive table bloat or database storage requirements, enabling more informed performance remediation and operational decisions.
Key Takeaway: Enterprise APIs and database metrics improve cloud governance, monitoring, and operational reliability.
7. Expanded Pipelines and Iceberg Data Integration
Mendix 11.14 makes Mendix Pipelines easier to configure by exposing Mendix-provided variables directly through the variable selector. Developers can use dynamic application, environment, and execution-specific identifiers when configuring POST and GET request steps, reducing the need to hardcode information when integrating deployment APIs and external enterprise systems.
The release also introduces an Iceberg bridge for Mendix Event Broker. The bridge transforms Change Data Capture events into Iceberg tables and propagates them into a data lake. Organizations using CDC with Mendix Event Broker can configure the bridge by selecting their data lake type and the CDC events they want to connect, enabling Mendix application data to be replicated into their data lake through a simplified integration process.
Key Takeaway: Expanded pipeline variables and Iceberg integration simplify automation and enterprise data connectivity.
8. Siemens Xcelerator – FDS Gateway Login Connector
Mendix 11.14 introduces the FDS Gateway Login Connector for integrating Mendix applications with Siemens Foundational Services. The connector enables authentication and Single Sign-On capabilities, allowing users to sign in using Siemens ID or a federated workforce identity provider while organizations continue using their existing identity and access management infrastructure.
By simplifying access to Siemens authentication services and standards, the connector helps organizations integrate Mendix applications more consistently within the Siemens Xcelerator ecosystem. This provides a more streamlined authentication experience while supporting enterprise identity management requirements across Siemens Xcelerator solutions.
Key Takeaway: The FDS Gateway Login Connector simplifies secure authentication and SSO for Siemens Xcelerator applications.
