ecu.test Release 2026.3

Highlights at a glance

ecu.test agent  chat

When creating test steps, a lot of contextual knowledge sits implicitly within the project and has to be taken into account manually for every single step. This takes time, creates inconsistencies, and is prone to errors.

With the ecu.test agent chat, AI support comes directly into ecu.test and addresses exactly that. It handles repetitive work when creating test steps, delivers consistent results, and answers questions about ecu.test. That frees up more time for what really matters: the actual test logic.

AI-powered test step generation

The agent assists in creating test case steps and generating trace analysis steps. In both cases, it automatically analyzes the existing project context, taking into account:

  • Signal mappings, tool and port jobs, and package mappings
  • Package variables and global constants
  • The block description as the basis for suitable test steps

Because the agent works directly with the entire project setup, the generated steps fit your existing test environment and are usually ready to use right away. This significantly reduces manual errors when selecting signals, mappings, or package calls.

Note: Support for trace analysis is still in an early stage of development and  will be expanded further in upcoming releases.

Consistency through reference implementations

The agent learns from similar blocks in existing packages. Proven test patterns from your own project are incorporated into the generation process and applied to new test cases. This results in consistent and standardized recurring solutions without requiring active intervention from teams.

Customizable generation

Generation can be flexibly adapted to project-specific requirements:

  • Configurable filters control how many elements from each category are included
  • Custom instructions allow for adaptation to project-specific naming conventions, style guidelines, and individual preferences

ecu.test help right in the chat

Not sure how a specific feature works or how to solve a problem? Just ask in the chat - the agent will help you without ever having to leave your environment.

Agentic Multimedia test step

With the new Agentic Multimedia test step, complex multimedia test sequences can be planned, executed, and evaluated autonomously based on a freely worded prompt.

The test step works with all multimedia tools available in ecu.test, as well as with custom user tools. The model used for multimedia processing is selected centrally in the workspace settings, under the Multimedia settings.

At runtime, the test step analyzes the prompt, derives suitable action sequences from it, and executes them on the linked multimedia tool (see example).

Optionally, previously determined prompt analyses, execution steps, and context information can be reused via a cache.

This allows repeated executions to run faster and reduces unnecessary model requests. The test step parameters let you control

  • whether the cache is used,
  • the maximum number of model requests allowed, and
  • how long to wait after each action.

With self-healing enabled, faulty cache entries can be discarded as needed, and the sequence can be re-executed without reuse.

Example:

The test step contains the following prompt: "navigate to the bluetooth settings and turn it off. In the end, check that you are in the bluetooth settings and that bluetooth is turned off."

The current screen content of the connected system under test (SUT) is captured and analyzed. The agent determines the next action and executes it step by step, until the goal is reached. 

In this case, it navigates fully autonomously to the Settings menu (step 1), switches to theConnected Devicesmenu (step 2), and disables Bluetooth(step 3) — without any hard-coded coordinates or manual test steps.

In the test report, each intermediate step is documented with the model's reasoning and a screenshot of the SUT.

Develop test cases locally, execute remotely via one:cx/ test.guide 

Test benches are a rare resource, and they deliver the most value when they're running tests almost continuously. Manually developing and adjusting new test cases, however, often ties up these resources and reduces overall utilization as a result.

With this feature, development and execution become one seamless workflow.

  • Develop locally, execute remotely: Test cases are created at your workstation and executed on the target test bench with just a few clicks.

  • Smart selection: In the execution dialog, you select the package changes, target configuration, and the right target system.

 

  • Automated queuing: Changes are automatically pushed via Git, and the test is added to the test bench queue.

  • Maximum utilization: The test bench stays free during development and is used exclusively for actual test execution – perfect for parallel use and CI automation.

 

Workspace compatibility between different ecu.test versions

Starting with ecu.test 2026.3, running different versions of ecu.test in parallel within the same workspace will be significantly easier. This is particularly helpful in environments where multiple teams share a workspace, such as library workspaces, and were previously forced to roll out updates simultaneously across the organization.

In the workspace settings, you can now specify an ecu.test version with which the artifacts in the workspace should remain compatible. When saving, ecu.test checks the artifacts against this target version, thereby preventing them from accidentally becoming incompatible with older versions. A dialog box alerts the user if an incompatibility cannot be avoided. Additionally, the Workspace Explorer directly displays which ecu.test version the artifact or library is compatible with.

Further additions:

  • A new migration wizard (see Help - Migration helper) conveniently updates an entire workspace to a new ecu.test version, eliminating the need to open and save each artifact individually.
  • The Object API has been updated accordingly so that saving packages via the API also takes the compatibility setting into account, making the feature fully usable in automated workflows.
  • Comprehensive user documentation covers setup, recommended procedures for version upgrades, handling shared workspaces, and limitations.

ecu.test Interface Python Library

The new ecu.testInterface Python Libraryprovides the public Python interfaces for developing custom ecu.test extensions. This makes it much easier to develop these extensions in an external IDE such as VS Code or PyCharm.

The library is provided as a Python wheel and includes API documentation as well as usage examples.

Modern IDEs enable autocompletion, including inline help. This also allows AI coding assistants such as GitHub Copilot or Claude to better support developers when creating, modifying, or validating ecu.test extensions.

Its use is described in detail in the ecu.test help documentation and currently includes the following API components:

You can find information about all new features, extensions, and improvements in this release in our Release Notes and in the Changelog.

ecu.test Release 2026.3

Changelog