TNG AI E2E TestGen

A Vision and Large Language Model-based agent

TNG AI E2E TestGen is a tool for automated and efficient end-to-end (E2E) test generation for web applications. Our approach utilizes a combination of Vision and Large Language Models to generate tests based on user-provided instructions in natural language (e.g. existing manual test descriptions).

How it works

Our tool uses artificial intelligence to generate clean, maintainable Playwright tests in Python, Java, C#, and TypeScript. TNG AI E2E TestGen runs locally, without requiring access to the application's source code, and involves a three-stage generation process:

Natural language execution: Users can simply describe their test intent in plain English (e.g., “Validate checkout flow”). Our AI agent automatically drives the browser, navigating and interacting with the application until the specified expectations are fulfilled.
Action-to-code refinement: The recorded interactions are then converted into Playwright tests following best practices. Users can provide a sample or style guide, and our tool will adapt the generated code to match their team's conventions.
Test execution and healing: Newly created tests are executed immediately and automatically refined or fixed in case they fail.

It is possible to use on-premise Vision and Large Languages Models or integrate with third-party APIs (e.g., OpenAI, DeepSeek, Anthropic, etc.), giving the choice to secure confidential projects or private tools.

Key benefits

Efficiency: Our tool helps in creating E2E tests more automatically, saving time and effort.
Convenience: TNG AI E2E TestGen is easy to set up and operates reliably on its own.
Consistent quality beyond demos: It produces consistent and reliable results, even in complex testing scenarios.
Data sovereignty: Runs locally on the user's machine within a standard Python environment and allows using self-hosted LLMs, thus keeping your data under your control.
Extensibility: The flexibility of our tool allows it to be extended to additional use-cases, such as documenting an existing web application.

Workflow in CLI

The screenshot shows a segment of an end-to-end test generation run with TNG AI E2E TestGen. In this scenario, the task is to find a specific speaker on the TNG Big Techday website. The agent has successfully loaded the homepage and clicked the “Speaker” navigation entry to open the speaker section. To achieve this, it combines DOM information with visual input from Playwright: it uses the rendered page to identify and colorize clickable UI elements and maps them back to their corresponding HTML elements in the DOM, enabling robust, visually guided interaction with the application.

You can see our tool in action in this demo video – from initial setup through execution to the finished test.

Interested?

To help you get started with TNG AI E2E TestGen, TNG offers guided access to the tool along with comprehensive support for its integration. Consulting services from our expert team can also be provided to ensure a smooth transition. We offer a personalized demonstration of our tooling tailored to your specific use case.