Our principles
- Evidence only. A rating reflects what a service actually does in a test, not its marketing.
- Independent. No viewer can buy a score; ratings follow the framework, not the payout.
- Safety first. When a service puts a visitor at risk (credential requests, private-access claims, deceptive funnels), that outweighs convenience.
- Public only. We assess tools for viewing public content; promising private access is penalised, never rewarded.
What the ratings mean
Each 0-100 score maps to a plain-language status that tells you what to do, not just a number:
- Safe (85-100): clean behaviour: no login, few trackers, clear policy. Reasonable for public stories.
- Low Risk (70-84): works well with minor caveats.
- Use with Caution (55-69): functional but notable: trackers, ads, or thin disclosure.
- High Risk (40-54): heavy tracking/redirects or misleading claims.
- Avoid (0-39): credential requests, survey funnels, or deceptive behaviour.
- Not Working: failed to return real results.
What we weigh
The score is weighted toward the signals that carry the most risk (credential safety, redirects, and trackers), with functionality, disclosure, and ads counting for the rest. The full point breakdown lives in the methodology, and every review shows how a service earned its score.
What we can & can’t assess
- We can observe login prompts, pop-ups, redirects, trackers, policy presence, account-access claims, ads, and whether results are real.
- We can’t see a service’s private backend or guarantee behaviour after our test date; sites change, which is why we retest monthly.
Editorial review
A tester records the evidence; a second reviewer checks it before a rating is published. Disputed or borderline cases get a second look, and material conflicts are disclosed. See the editorial policy.
When ratings change
Scores are retested every cycle and after any significant change to a service. A substantiated correction triggers a retest, and past scores stay visible in the monthly reports.