About CompanionRank

Why this site exists

Every AI companion platform describes itself in almost identical language. The most realistic. The best memory. Unlimited everything. Search for a recommendation and you mostly find listicles that rank whichever product pays the most, written by people who never opened the app.

We started CompanionRank because that was the whole market. There was no source that told you which company you were actually paying, what its refund terms said, or whether the memory feature it advertised was real. So we built one.

How we are different

• We read the code, not the landing page. Where possible we inspect the platform's own shipped client software and public API responses to confirm what it can actually do.
• We check the corporate paperwork. We name the operating company and its registration number, pulled from the relevant national registry, for every platform we cover.
• We read the contract, not the advert. We have found platforms advertising "lifetime" benefits whose own terms cap them at six months. Where the two disagree, we score the contract.
• We label what we could not verify. Claims we confirmed ourselves are marked Verified. Claims that rest on the operator's word alone are marked Reported. We would rather admit the gap than dress it up as testing.

How we make money

Some links on this site earn us a commission when a reader subscribes. That is how the site is funded, and we set it out in full in our affiliate disclosure.

Commission rate is not an input to our scoring, and no platform has ever been offered the chance to buy a better position. The proof is in the table: the lowest-scoring platform we cover pays us considerably more per signup than the highest-scoring one does. If we ranked by what paid best, the list would be upside down.

Our scoring in brief

Every platform is scored out of 10 across 6 weighted categories: Conversation, Memory, Media, Voice, Value, Trust. The weights are fixed, applied identically to every platform, and published on our methodology page so you can check our working.

What we cover

We currently benchmark 6 platforms: Secrets AI, Candy AI, Swipey AI, DreamGF, Nomi AI, CrushOn AI. We add more as they become significant enough to matter, and we re-score existing ones when their pricing or features change. Last full revision: 2026-08-22.

Corrections

We get things wrong sometimes, and prices in this category move fast. If you can show that something here is inaccurate, tell us and we will correct it and note the change. That applies equally to operators who think we have scored them unfairly.

Editorial independence

No platform reviews our content before publication. No platform has approval over its score. We have never removed criticism in exchange for a higher commission rate, a payment, or anything else, and we will not.