Reader question first
Every page begins with a defined reader job. The home page explains the category, the Reviews hub compares five products, individual reviews answer service-specific decisions, and Guides teach repeatable workflows. Closely related keywords are assigned to one canonical page when they express the same need. We do not publish a new page only to repeat a phrase.
Products are selected when they represent a distinct workflow and enough current material exists for a useful review. A service can be omitted when its product status, pricing or adult-content boundary cannot be described responsibly. Inclusion is editorial coverage, not certification.
Evidence before verdict
Mutable claims should come from current official product, pricing, help, checkout or policy material. Direct interface observation may support what a public route shows. We record the checked date and separate an official statement from our interpretation. Conflicting statements are preserved as uncertainty rather than converted into false precision.
A documented feature does not prove output quality. We do not imply a paid test, repeated prompt benchmark or account-level inspection unless a reproducible record exists. Promotional galleries inform the description of positioning but do not become evidence of consistency.
Writing and scoring
Each review covers the starting action, continuity tools, image and video path, US-dollar pricing, free access, best-fit user and decisive limitation. Annual offers show the amount charged as well as the monthly equivalent. Checkout-dependent prices are labeled, and stale promotional numbers are not substituted for missing evidence.
Seven weighted criteria produce the visible score. Missing performance evidence receives a neutral or cautious treatment rather than an optimistic estimate. Scores help order a shortlist; the prose explains whether that order matches a specific reader.
Editorial independence
Every product follows the same source requirements, scoring weights and limitation rules. A reviewed company cannot buy edits, suppress criticism or approve the final verdict. Logos are used for identification and commentary, while supplied AI-generated models and videos remain editorial media rather than product-output evidence.
Updates and corrections
Prices receive a defined freshness window and should be checked again when that window expires or a credible correction arrives. Product availability, media features and policy changes may require earlier review. A material update changes the visible date and every synchronized format affected by the claim.
Readers, rights holders and product teams can submit a factual correction through corrections. We compare the report with the current source and existing record. Correct evidence can change the page; pressure or brand preference cannot.