Not what the ads say. Every tool runs through nine real-world sources — Reddit, YouTube, Hacker News, GitHub, package downloads, review platforms, momentum, pricing value, editorial tests — each worth fixed points, summed into a 100-point verdict. Expand any row on any list and check every source yourself.
First run — movers appear tomorrow when today's ranks become yesterday's baseline.
Because you can check our math — source by source. Every ranking row expands into the nine source cards behind it: points per source, raw metrics, and links to the actual Reddit threads, HN stories, repos, packages and review pages. The 100-point rubric is published and generated from the same config the agents compute with. We take no sponsored placements, and affiliate terms are banned from scoring. When a source fails, it shows 0 and says so — nothing is invented.