Editorial Policy
What We Publish
AI Girlfriend Lab covers AI girlfriend apps, character chat platforms, and adjacent companion products. We publish reviews, comparisons, best-of lists, buying guides, and pricing explainers. Topic selection starts with reader search intent and competitive gaps—not vendor press releases. Batch-1 (July 2026) focuses on nine core competitors plus trust, pricing, and choose guides so new readers land on a complete cluster rather than orphan posts.
How We Choose Topics
We prioritize pages that answer a decision: “Is X worth it?”, “X vs Y?”, “Best apps for Z?”, and “How much does it cost?” Trust pages publish first so later reviews can link to a real methodology. We add products to the database only after a minimum test plan is defined—usually five or more hours across desktop and mobile.
What Comes From Testing vs. Official Sites
From hands-on testing: conversation behavior, memory probes, UI friction, media output quality, and time-to-first-good-chat. From official sites: pricing tables, feature lists, and policy URLs—we always cite the check date. Editorial judgment: who should choose or avoid a product, category awards, and comparison winners when scores are close. We never invent user counts, star ratings, or quotes from anonymous users.
Independence and Rankings
Product rankings are based on our published methodology, not on affiliate commission rates. Sponsored placements, if any, are labeled separately from editorial rankings. Affiliate partners—including AISOUL where linked—do not automatically win every category.
How We Verify Prices and Features
Editors open official pricing and checkout pages in a clean browser session. We record auto-renewal language, trial terms, and currency. When a feature is beta or regional, we downgrade the testing completeness percentage and exclude it from scoring until we can verify it.
Corrections and Brand Outreach
Readers and vendors can submit corrections through our corrections form. We verify claims against reproducible steps or primary sources. Verified errors are fixed with an update log entry; significant mistakes may include a short correction note at the top of the article. Brands may not purchase ranking position or rewrite conclusions. If a vendor disputes a score, we re-run the relevant portion of the test plan on a fresh account when feasible and publish what changed.
Mature Content Standards
We describe mature capabilities informatively for adult readers. Public pages use non-explicit screenshots and crops. We do not embed explicit imagery in reviews. Products with age gates are noted in pros/cons and privacy sections.
How Often Articles Change
Active competitors in our batch-1 cluster are reviewed on a 60-day feature cycle and 30-day pricing cycle. Evergreen guides update when market pricing shifts materially. Each article shows tested and pricing-verified dates where applicable.