Editorial policy

How we compare AI companion services.

Current methodology: 26 September 2026

Sources

We prioritize providers’ official product pages, policies, help centers, pricing and checkout disclosures. We distinguish provider claims from our own observations and avoid treating affiliate landing-page claims as independent evidence.

Comparison dimensions

No synthetic scores

We do not compress unlike products into a single numerical rating. “Best suited to” labels describe a use case; they do not claim that one service is universally better.

Corrections and updates

Feature and pricing claims are date-stamped and periodically reviewed. Providers and readers may submit a correction with a supporting official URL through the contact page. Commercial relationships do not prevent corrections or negative findings.