What's new

OpenText™ Core Performance Engineering Aviator is a cloud-based enterprise service that delivers Artificial Intelligence (AI) capabilities. The service extends functionality within OpenText Performance Engineering products that are connected to the paid Aviator service. For details on supported client instances and integrating with Aviator, see the OpenText Core Performance Engineering Help Center.

Aviator provides scripting support for Performance Engineering products through VuGen, and automated, AI‑assisted analysis through Core Performance Engineering Analysis.

Aviator for Scripting

Aviator for Scripting includes the following enhancements:

• Ability to use single sign-on from your browser to complete Aviator authentication.

• Ability to create new data files and add them under the script folder.

• A descriptive summary of the parameters used in the script, provided on request.

• Context-aware scripting assistance that can suggest next steps and relevant actions, based on script content and execution context.

• Advanced language model upgrade, enabling more detailed, accurate, and comprehensive responses.

• Extended assistance with error logs for Java protocols.

• AI assistance available in the TruClient 2.0 - Web JavaScript editor. For details, see Aviator AI actions in the JavaScript editor.

• Stability and optimization improvements.

For details, see Aviator for Scripting.

Aviator for Analysis

Aviator for Analysis includes the following enhancement:

• Ability to use predefined prompts, and append results to the current conversation, without losing context.

For details, see Aviator for Analysis in the OpenText Core Performance Engineering Analysis Help Center.

General Aviator updates

The following enhancements are available for both Aviator for Scripting and Aviator for Analysis:

• AI Usage report, available on the AI Usage page of the Core Performance Engineering Platform, providing aggregated visibility into AI query consumption across supported products.

• Chat management, enabling you to reopen, pin, and rename conversations in the history.

• Improved functionality and responsiveness.

MCP support

Model Context Protocol (MCP) provides a standardized and secure way to leverage AI-driven natural-language prompts, enabling the streamlining and automation of performance engineering workflows.

MCP server for Controller Service

The new Controller Service MCP server enables you to initiate Controller Service processes through external AI tools. You can integrate the MCP server with your favorite AI client, and then use the server tools in chat to run and stop load tests, view Vuser details, and more.

For details, see MCP integration in the OpenText Professional Performance Engineering Help Center.

Note: MCP server for Controller Service is a private Beta release.

MCP server for DevWeb

The DevWeb MCP server includes the following enhancements:

• Improved handling of the record, generate, and replay processes.

• Expanded guidance for working with the MCP server workflows and processes.

• Provides assistance with writing scripts.

• Ability to identify high-confidence parameterization candidates, such as user names, passwords, and user identifiers (Beta feature).

• Ability to create DevWeb scripts using Swagger API specifications, and to edit function parameters.

For details, see MCP integration in the OpenText Performance Engineering for Developers Help Center.

A new Blazor - Web protocol is available for recording and replaying Blazor applications with decoded SignalR MessagePack payloads.

For details, see Blazor - Web protocol.

Support added for Citrix Workspace/Citrix Virtual Apps and Desktops version 2507 LTSR.

For details, see Citrix ICA protocol.

You can now specify an IP version to use during script execution, forcing the system to work with IPv4 or IPv6 only.

For details, see DevWeb runtime settings.

Code generation for Kafka now uses the simplified Builder pattern, replacing the XML raw deserialization that was used previously in the code.

For details, see Code generation.

• Support added for Anthropic model.
• Support added for custom LLM models that are non-streaming, JSON-formatted, and include token counts.
• Ability to configure cross-step execution, controlling whether the script waits for the LLM request to complete.

For details, see LLM protocol.

• If TruClient shuts down unexpectedly, you can recover a script from its last known state. For details, see Recover a script after a crash.

• You can now call a Javascript function using a Generic API action step, and create a variable if needed. For details, see Generic api action.

• Support added for two new client-side measurements, Cumulative Layout Shift (CLS) and Interaction to Next Paint (INP). For details, see Performance metrics.

• AI assistance is now available in the JavaScript editor for code generation and code-fix suggestions. For details, see Aviator AI actions in the JavaScript editor.

• Support added for speed simulation on Windows and Linux.

• UI improvements for all code editors.

• Security fixes.

For details, see TruClient 2.0 - Web protocol.

• Support for Chromium browser version 149.

• Security fixes.

For details, see TruClient - Web protocol.

• Enhanced web-recording performance.

• Improved recording performance and certificate creation when recording with proxy.

• Non-blocking socket support.

• Improved WebSocket recording.

For details, see Web - HTTP/HTML protocol.

Support added for Windows 11 version 25H2.