Independent dating app reviews 24 files open · facts last checked 22 NOV 2026

Method

How a file gets built.

Every app goes through the same three stages: the paper trail, the field test, the verdict. A file can be published after stage one, but it only gets a score after stage three.

Last updated November 19, 2026

The paper trail

Every file starts with the documents an app would rather you skimmed. We read the current pricing pages and in-app prices, the privacy policy and terms of service, the app store listings and privacy labels, the owner's earnings reports and filings, and any regulator decisions, court cases and settlements, plus serious reporting on the company.

Every factual claim gets a footnote linking to its source. Every file shows the date its facts were last checked. Prices for dating apps change often and vary by country, age and platform, so we always say which price we quote and when we saw it.

A file at this stage is labeled Fact file · No score yet. It tells you what the app is, what it costs and what it does with your data. It doesn't tell you what it's like to use, so it doesn't pretend to.

The field test

We then use the app on a real account, with real details. Each review gives the dates we tested, the plan we used and anything still pending when we published. We:

  • go through sign-up, onboarding and verification, noting every permission and data request along the way;
  • use the free tier until we hit its limits, and record when and how often the app pushes us to pay;
  • pay for a subscription, or use a premium account the company gives us, which the file then says. We note which plan we used and check what it actually changes compared with the free version;
  • test the safety tools: verification, reporting, blocking, hidden-words filters and safety centers;
  • file a data access request and look at what the app holds on us — the file it keeps on you. An answer can take up to a month, so if it arrives after we publish, we add it to the review with the date;
  • cancel the subscription and delete the account, timing how many steps and screens it takes. If deleting has to wait for the data request, we add the result later and say so.

What we won't do: create fake profiles, mislead other users, or run "experiments" on people who are there to date. Other people's faces, names and messages are blurred or cropped out of every screenshot we publish. Screenshots on fact files come from public pages, viewed with no account signed in, and each one shows its source and capture date. Yellow highlights are ours. On adult services, fact-file screenshots are taken with the page's images blocked. Field-test screenshots show the characters as the service shows them to a signed-in user, but never nudity or sex acts.

For AI companions we also test whether it remembers what we told it in an earlier session, how hard the app pushes paid features, and how it responds when a user talks about loneliness or self-harm.

The verdict

After the field test, the app is scored out of 10 on five pillars. The overall score is a weighted average:

PillarDating appsAI companionsWhat it measures
Value20%16%What the free tier really allows, and what paying actually gets you.
Experience25%36%Onboarding, design, the quality and activity of the people you are shown, and how the app behaves day to day. For AI companions: the range of characters and features, chat quality, memory, and image and video quality.
Safety20%16%Verification, reporting and blocking, scam and abuse protection — and for AI companions, how the app handles vulnerable users.
Privacy20%16%What data is collected, who it is shared with, what our own data request turns up, and how hard it is to delete.
Honesty15%16%Dark patterns, hidden prices, pressure upsells, fake-looking activity and how easy it is to cancel.

AI companions are weighted differently. With a dating app, the product is a way to meet other people. With an AI companion, the product is the whole experience: the characters, the chat, the memory, the images and video. So Experience counts for 36% there, and the other four pillars count equally. We made this change on November 19, 2026, and it applies to every AI companion we score.

How to read a pillar score. A 5 means typical for the category: what most of the services we've researched do. A 7 or 8 means clearly better than most, and a 9 or 10 the best we've seen. A 3 means clearly worse than most, and 0 to 2 means it causes real harm. For AI companions we compare with the 36 services we had researched by November 19, 2026, and we update that comparison as we cover more. Where we know too little about the others to say what's typical, the review says so.

You might notice something missing: a "success rate". We don't score whether an app will find you a partner, because nobody outside the company can measure that honestly. Claims like "a match every few minutes" are marketing, and when we quote them we label them as the company's own claims.

How we make money

We earn money from advertising links, also called affiliate links. If you click one and buy, the company pays us a commission, at no extra cost to you. Our current programs are: Secrets AI, Promptchan, aiAllure, OurDream, Xeve, Swipey AI. There are no display ads on this site. We usually pay for the subscriptions we test. When a company gives us free premium access instead, the file says so.

Every link that earns us money is labeled where it appears, and pages that contain them say so at the top. Which apps we cover and test first is partly a business decision, and we favor apps we can earn from. What we report about them isn't for sale: a commercial relationship never changes a fact, a score or the problems a file lists. No score or placement can be bought. If an app ever gives us free premium access, the file will say so.

Corrections

Dating apps change constantly, and we will sometimes be wrong or out of date. If you spot an error, email hello@datingfile.com. We fix confirmed errors quickly and note significant corrections at the bottom of the file, with the date.