What test types can it build prompts for?

Three. Unit test cases for a specific function or component, an integration test plan covering how components interact, and an exploratory testing charter that guides manual session-based testing. Pick the one that fits your stage.

How does risk level change the output?

Higher risk asks the model for deeper edge-case coverage, negative tests, security and concurrency cases, and explicit failure modes. Low risk keeps it to the core happy and unhappy paths so you are not over-testing throwaway code.

Will it generate runnable test code?

For the unit test type, yes — the prompt asks the model to emit complete test files in your chosen framework. Integration and exploratory outputs are plans and charters, which are intentionally prose so a human drives them.

Do I paste my source code?

Yes, paste the function, component, or API under test after the prompt. The model uses your real signatures and behavior to write specific assertions instead of generic placeholders.

Which frameworks are supported?

The prompt adapts to whichever framework you choose — Jest, Vitest, pytest, JUnit, Go testing, and others — by instructing the model to use that framework's idioms, assertions, and mocking style.

Testing Strategy Prompt Builder

Testing strategy prompt builder

A useful test suite starts with a clear charter: what are we testing, how badly could it break, and what kind of testing answers that. This builder captures those decisions and turns them into a precise prompt so an LLM produces runnable unit tests, an integration test plan, or an exploratory charter that actually fits your feature and stack — not generic boilerplate.

How it works

You name the feature, set its risk level, choose your framework, and pick a test type. The tool assembles a prompt that instructs the model to scale coverage to the risk, use your framework’s idioms, and produce the right artifact: complete test files for unit testing, a structured plan with scenarios and data for integration testing, or a session-based charter for exploratory testing. Paste the prompt plus your source code into any LLM. All generation happens locally in your browser.

Tips and examples

Match risk to reality. A payment flow is high risk and deserves negative, concurrency, and security cases; a tooltip is low risk and does not.
Paste the real signature. Giving the model the actual function or API lets it assert on real return shapes instead of guessing.
Use exploratory charters for new features. When behavior is still fluid, a charter guides a focused manual session better than brittle automated tests.
Combine types over time. Start with a unit-test prompt, then run an integration-plan prompt once the pieces connect.