How the Trust Score works
A plain 0–100 score that shows how openly an AI tool explains who runs it, what it costs and what it does with your data. Every point comes from the tool’s own published policies, and every source is linked.
What we check
| Check | Points | What earns the points |
|---|---|---|
| Company | 20 | Company name, country and a working contact are published (10 points if only the name is) |
| Training on your data | 20 | 20 if the policy says your content is not used to train AI; 10 if it is used but you can opt out; 0 if it is used with no opt-out or the policy doesn’t say |
| Pricing | 15 | Prices are published on the website |
| Data deletion | 10 | You can delete your data or account yourself |
| Data retention | 10 | The policy says how long your data is kept |
| Security | 10 | An independent certification is published, such as SOC 2 or ISO 27001 |
| Refund and cancellation | 10 | Refund or cancellation terms are published |
| Free option | 5 | A free plan or free trial exists |
Grades: A = 85–100 · B = 70–84 · C = 50–69 · D = below 50.
Rules we follow
- We only use what the tool publishes itself: privacy policy, terms, help centre, trust or security pages, pricing page.
- If something isn’t stated, it scores zero. We don’t guess. Items we haven’t reviewed yet are labelled “not reviewed yet”.
- Every score shows the date we checked and links to its sources. Policies change, so we re-check regularly.
- The score measures transparency and data practices. It is not a rating of how good the tool is at its job.
- No company can pay for a higher score. Companies can ask us to re-check after they update their policies.
Spotted an error?
If a policy has changed or we have misread it, tell us with a link to the source and we’ll re-check it.