Our question is narrow on purpose: does this product do what it claims, for the kind of person it’s sold to?
1. Claims come from the product, not from us
We collect specific claims from the product’s own website, pricing page, docs and launch announcements, and quote them word for word.
2. We test on a real plan
We use the plan most buyers would choose. If we test on a free trial or a vendor-provided account, the review says so.
3. Same tasks, same inputs
When we compare tools, each gets identical tasks, files and prompts — published with the review where possible.
4. We record the real cost
List price, actual price after add-ons and usage limits, and what happens when you hit the cap.
5. One label per claim
No single score that blends a great feature with a broken one. Each claim gets its own verdict:
| Label | Meaning |
|---|---|
| Verified | We tested it and the claim holds as stated. |
| Holds, with caveats | Mostly true, with a limit you should know about. |
| Overstated | Something real underneath, but the marketing promises more. |
| Doesn’t hold | Our tests contradict the claim. |
| Untested | We couldn’t test it yet, and we say why. |
6. Evidence cutoff and change log
Every review states when it was tested. If the product ships changes that affect a claim, we re-test and log it.
7. Corrections
If we get something wrong, we fix it in the open and record the fix.
8. Money and independence
Affiliate links help fund testing. They don’t affect labels. Vendors don’t see reviews before publication. Full disclosure policy →