Methodology

Weights

The same five weights apply on every page unless the content file records a reason for a change. This baseline does not change them.

• 40%Task qualitySame prompts for every tool. Not run yet.
• 20%Price / valueEntry price. Unknown prices score a neutral 50.
• 15%Ease / speedHow fast a fluent person gets a usable result.
• 10%EcosystemIntegrations and the tools around it.
• 15%External signalsPublic benchmarks, only with a link and a date. None attached yet.

quality 40% · value 20% · ease 15% · ecosystem 10% · signals 15%

Rules

• Affiliate programs and sponsored placements are not score inputs. The scoring function does not read them.
• A sponsored slot, when one is sold, sits outside the ranking and carries a visible Sponsored label.
• Unknown prices are marked TODO. Their value subscore is a neutral 50, not a guess that the tool is cheap or expensive.
• Quality is editorial until the shared prompt suite is run and the outputs are published on the page.
• External signals need a public benchmark URL and a date. None are attached in this baseline, so that 15% is also editorial.
• A human must sign changes to the top 3. humanApproved is false on every use case right now.
• Rank changes go in content/changelog.json with a date and a reason.

Tie break

Higher total score wins. If totals match, higher task quality wins, then higher value, then the tool id in alphabetical order.

Changelog

• Oct 14, 2026 · Published the first editorial baseline for 10 use cases.

  MVP seed. The prompt suite has not been run. A human has not signed the top 3. Prices marked TODO use a neutral value score of 50.