Trustworthy AI TestingKit
Sign in

Try it · no sign-in

Would you sign this off?

Sample AI summary

Executed: 152 · Passed: 142 · Pass rate: 95%
All functionality verified.
Risk: Low. Release recommended.
TC-14 Expected result: System behaves correctly.

Sample source · tester decision: No-go

Executed: 152. Passed: 142. Failed: 10. TC-14: On invalid input, display the validation error. Only the reset workflow was tested. Release decision: No-go.

Your AI · v1.0 · for QA engineers

AI drafts.
You verify.

Works with the AI you already use. Nothing is sent from this app.

The app never calls AI. Private, safe-checked account sync is separate.

A working bench for testers who use LLMs and need to show their work can be trusted: fill-in prompts with a built-in data leak scanner, an interactive review checklist, and an evidence log you can hand to an auditor.

Open the workbench New here? Read this firstReview an AI output

// the principle

  • → The AI is a drafting assistant, not an oracle.
  • → You own the result. “The AI wrote it” is not a defence.
  • → Verification is the job. Reinvest saved time in checking.
  • → Evidence beats assurance. Turn “trust me” into something reviewable.

38

prompts

15

checks

10

sections

The workflow

  1. 01

    Prepare

    Remove or replace anything confidential. Gather the real requirement.

  2. 02

    Prompt

    Fill every [PLACEHOLDER], copy into your own AI and paste its answer back.

  3. 03

    Review

    Read the ‘What to check’ note, then run the 15-point checklist.

  4. 04

    Correct

    Edit, delete or add. Re-prompt if it has gone badly wrong.

  5. 05

    Record

    Log what you shared, changed and verified — for anything that matters.

Prompt Workbench

Fill placeholders in a form, see the live prompt, and get warned before you paste an email, card number or API key.

Review Checklist

Tick, cross or N/A all 15 checks. Three or more fails? The kit tells you to re-prompt instead.

Evidence Log

Record task, prompt, data shared, changes and verification. Export as Markdown or CSV.

Prompt library

1 · 4 prompts

Test case generation from requirements and user stories

2 · 4 prompts

Edge cases and negative testing

3 · 4 prompts

Bug report writing

4 · 4 prompts

Risk analysis and test strategy

5 · 4 prompts

Exploratory testing charters

6 · 4 prompts

Test data generation

7 · 4 prompts

Automation help (Playwright / Cypress)

8 · 3 prompts

Regression test selection

9 · 4 prompts

Accessibility checks (WCAG 2.2)

10 · 3 prompts

Release sign-off summaries

Never paste confidential data. If you'd hesitate to email it to someone outside the company, don't paste it into an AI tool. Read the short rules →