Readable by humans
Express expected behavior in clear, natural-language test steps anyone can review.
The specification layer for agentic coding
Don’t chat with your code. Specify it.
AI agents can generate software in minutes. TestWeasel gives them something better than another prompt: a reliable, executable definition of correct behavior.
Free for Windows · Your model · Your Playwright code
Agentic coding changes everything
Code is abundant.
Clarity is not.
AI agents can produce thousands of lines of code before a human finishes a review. But speed does not answer the essential question: what is the software supposed to do? The faster implementation becomes, the more valuable a precise specification becomes.
A new software primitive
The specification is the product. The code is just a build artifact.
Code can be regenerated, refactored, rewritten and optimized. Tomorrow’s software will be built again and again by AI.
What must remain stable is human intent: understandable, reviewable and executable for every agent that touches the product.
Test First. Still Right.
Agents need a definition of done.
Agentic coding does not make test-first development obsolete. It makes it essential. Without an executable specification, an AI agent has nothing reliable to optimize against. Without tests, every iteration is another guess.
Give the agent a target it can prove—not a prompt it can interpret.
Executable by machines
Turn intent into robust Playwright tests that verify the real application.
Stable through change
Let agents rewrite the implementation while your specification stays in control.
What TestWeasel does
Human intent in.
Executable certainty out.
Describe behavior
Write what a user should be able to do in natural language.
Identify the UI
TestWeasel finds the right interface elements with intention-aware AI.
Generate Playwright
Get portable, maintainable test code in the ecosystem you already trust.
Let agents iterate
Change the implementation as often as needed—and prove it still works.
Living documentation
Generate your living documentation.
Every test case automatically extends your user docs. See the Living Documentation of TestWeasel
A better basis for review
Review with context.
Short enough to understandA behavioral specification captures intent without drowning it in implementation detail.
Clear enough to approveDevelopers, QA and business stakeholders share one readable source of truth.
Precise enough to testTurn approved requirements into Playwright tests that check the software’s behavior.
Durable enough to versionYour specification evolves with the product and stays available alongside the code.
Human Intent. Machine Execution.
Humans own the what.
AI owns the how.
People decide which outcomes matter. Agents decide how to implement them. TestWeasel connects both worlds with a specification that humans can trust and machines can run.
AI generates code. TestWeasel generates certainty.
Built for agentic development
Your agent needs a reliable oracle.
Whatever writes your code today—or tomorrow—needs an independent definition of correct behavior. TestWeasel works alongside your coding workflow, not inside a vendor lock-in.
CodexClaude CodeCursorGitHub CopilotYour next coding agentPlaywright
You’re excited about Agentic Coding.
But did you forget Test First?
AI speed. Human control.
Start with the specification.
Build software AI can change safely. TestWeasel is free to use, and the generated Playwright code is yours.
Windows available now · Linux and macOS coming soon