How we test

Every test page on didyoutryai is built from a saved test log. If we didn't run the test, the page doesn't exist.

The method

1. Pick a real task people actually search for, like writing an eBay listing or a contractor's follow-up email.
2. Write realistic inputs: the facts a real person would have. Where real customer data would be needed, we use clearly labeled fictional examples.
3. Run the same prompt on 2 to 4 tools and save every output exactly as produced.
4. Score each output against a fixed rubric for that kind of task (for example accuracy, tone, ready to use). Scores are spot-checked by a human editor.
5. Fact-check every price and feature claim against the tool's own site.
6. Re-test fast-changing tools every 90 days and show the date on the page.

How AI is used here

We use AI agents to run tests, draft pages and check facts. A human editor sets the test designs, reviews samples of every batch and owns the final call. We say this openly because hiding it would undercut the point of the site.

What we won't do

• Rank a tool higher because it pays a commission.
• Publish a page without a saved test behind it.
• Sell scores. Sponsored tests are labeled and scored the same way.