Methodology and ranking algorithm

1. Data sources

The initial dataset was compiled by hand and is continuously extended through two paths: user submissions (reviewed by AI and then a human) and editorial verification. Every entry carries a trust level:

• Unverified: from publicly compiled seed data, not yet individually confirmed
• Verified: a human confirmed the site's ownership and accuracy
• Official: official sites of governments, public bodies or platforms

2. Ranking algorithm

Rankings combine the following factors:

• On-site clicks (time-decayed, with the last 7 days weighted highest)
• Trust bonus (verified and official entries receive a fixed bonus)
• Information completeness (description, icon, and so on)

To avoid the top of the list becoming permanently fixed, click counts are log-scaled and newly added sites receive an initial exposure period. Repeated clicks from the same source are de-duplicated to reduce manipulation.

Please note: these rankings reflect behaviour on this platform and do not represent third-party traffic figures. We neither use nor claim to use any third-party traffic data.

3. Submission review process

1. Human verification: filters out automated bulk submissions
2. Automated pre-checks: validates link format and blocks known prohibited categories and obvious spam
3. AI review: scores URL consistency, promotional patterns, prohibited content, information quality and category fit
4. Human confirmation: a person decides whether to list the site and corrects the name and category

Submissions scoring below the threshold are rejected; those passing AI screening enter the human queue; anything ambiguous is always routed to a human. AI never decides publication on its own.

4. How we handle mistakes

If you spot inaccurate information, please report it via the submission page. Content that should not be listed will be withdrawn.