V 1.2.5

AI agents and MCP-compatible AI tools can now connect to your APIs directly, with enterprise-grade access control built in:

Turn your APIs into MCP servers in a few clicks — no separate infrastructure to stand up or maintain.
Access controls and data scoping let you decide exactly which capabilities an AI agent can use and which data it's allowed to see.
Reusable, permissioned prompts — author prompts once and control who's allowed to use them.
One-click client setup — every MCP server comes with ready-to-copy connection instructions and a built-in way to test your access token, so connecting a new AI tool takes minutes, not a support ticket.
The same security rules everywhere — the access and authorization rules you set for your APIs automatically apply to your MCP servers too, so there's nothing extra to configure.
Visibility into usage — see which AI tools are connecting to your MCP servers and how often.
A number of usability improvements to make setting up an MCP server simpler and clearer.

A new, more granular way to control access to your APIs:

Define reusable security profiles that control what scope of access is granted and what data is returned for a given token.
Require specific user or system claims before an API call is allowed to proceed.
These rules are enforced consistently everywhere the API is used, with no extra setup required downstream.

The AI assistant can now be opened directly from the content you're working on, and it stays open as you move between pages instead of resetting.

Chat now supports invoking published MCP servers, allowing users to access approved tools and capabilities for both internal and external client use cases.

NxAgent now supports publishing, deploying, and executing workflows, extending its capabilities across the workflow lifecycle.

The Analytics tab is now integrated into the Metrics dashboard, providing a more consistent and consolidated experience for accessing platform metrics and usage insights.