HOW THE DIRECTORY WORKS
Evidence first.
Unknowns included.
A directory should help you ask better questions. We link the sources behind each profile, distinguish AI claims, and show where the research is incomplete.
What the score means
The research evidence score runs from 0 to 100. It measures how well a profile is supported by attributable documentation across five areas.
It is not a product-quality score, a hands-on test, a user-review average, a safety assessment, or a forecast of investment returns. A documented product can still be unsuitable for your needs. A low score may simply mean we have not found enough evidence.
Feature documentation
Can we trace what the product does to substantive product pages or documentation?
Pricing clarity
Is the pricing model explained by a current, attributable source?
Maintenance & activity
Do dated releases, updates, or other sources document recent product activity?
Company & team
Can we verify the company or the people behind the product?
Launch history
Is the product launch or founding timeline supported by a dated source?
How points are awarded
Each published rating includes a component-by-component rationale and a scoring date. Version 1 awards the full category weight when qualifying primary evidence is recorded, or zero when it is missing. A reachable URL alone does not qualify: the source must substantiate the category. This measures documented coverage; it does not compare product effectiveness. The displayed score is the sum of the component scores.
Methodology version: evidence-coverage-v1. Scores can change when sources, products, or research change. They are not pay-to-rank placements.
A living source record
Sources are checked at a point in time. Vendor documentation establishes what a vendor claims; it does not independently prove effectiveness. Before choosing a tool, read its current pricing, availability, terms, and risk disclosures.
Explore Prediction market tools