A directory that quietly ranks its own products is running a con. So here is the whole thing on one page: which sites are ours, how they are treated everywhere they appear on this site, and exactly what they can and cannot do as of July 2026.
The people who run EROTIC.CHAT also build and operate latenight.chat, goodboys.chat, goodgirl.chat and youup.chat. They are our own products. Wherever one of them appears anywhere on this site, it appears in a separately-labelled block marked Our Site, below the ranked list, with no score, no star rating and no rank number — because we are not in a position to rate our own work and you would be right not to believe us if we did.
No score. No stars. No position in a Top 5. They sit in their own block, clearly separated, clearly labelled. If you ever see one of our sites carrying a number in a ranked list on this site, that is a bug and you should tell us.
Everything we say about our own sites on this site has to be something you can confirm yourself in about a minute: load the page, read the published copy, view the source. Design decisions and published policies qualify. “It’s better” does not.
No user counts. No download figures. No testimonials. No review scores. No “trusted by” badges. None of these sites has a user base yet, and we are not going to pretend otherwise in order to look established.
They are new. They have no track record. Payments do not exist on any of them, so nothing is purchasable and no premium tier can be bought. And the part most disclosure pages would leave out: the conversation itself is not publicly reachable yet. The landing pages, published policies, pricing stances and the goal tracker on goodboys are live and real; the chat engine behind each site is not open to the public as of July 2026. You can verify this from the page source — the engine address on each site is either unset or pointed at a local machine. We would rather tell you that than let you find it by clicking.
Because the failure mode in this category is not bad products, it is bad information. The affiliate review layer and the actual users are describing different products — one app runs a 4.3-out-of-5 editorial average against roughly 2.1 from its users. A directory that ranks its own sites into the top slot is just the same disease with an extra step. The only version of this that is worth anything is the version where we tell you what we own.
Listed alphabetically. Not ordered by preference, and not ordered against anything else on this site.
Three named characters — a bartender working the closing shift, an overnight radio host, an insomniac photographer — instead of a library to browse.
What you can check yourself:
What’s genuinely different: Persona-first rather than catalogue-first. Every large product in this category leads with volume — 300,000 characters, 7,500 characters, a browse grid. This one ships three and bets that a specific written person outperforms a big grid.
The honest caveat: Unproven. Three characters is also just fewer characters, and if the writing is not exceptional that is a worse product, not a differentiated one. Nothing here has been tested against users.
Appears on: AI Porn Chat · AI Boyfriend
An accountability product with an adult register — you set real goals, check in, and the character’s warmth is a consequence of whether you showed up.
What you can check yourself:
What’s genuinely different: It is the only thing in this market we found that fuses a utility loop with the adult register. Every product surveyed retains users through content volume or a message cap resetting; this one retains through whether you went to the gym. The progression is also stored server-side state rather than inferred from a context window — which, if it holds up, would make it structurally immune to the failure mode that produced three separate 2026 incidents elsewhere (a model gets swapped, and the relationship resets).
The honest caveat: Also the riskiest thing in the network. A product that invites photo check-ins would be holding the most sensitive data anywhere on this site, and the ephemeral-by-default policy has to ship with that feature rather than after it. We are saying that here, in public, before the feature exists.
Appears on: AI Boyfriend
Four characters, and a bet that the memory is the product rather than a premium add-on.
What you can check yourself:
What’s genuinely different: Two things, and both are direct answers to failures we documented in the reviews on this site. The first is a memory page you can open — see what it wrote down, edit what it got wrong, pin what matters, delete what you would rather it forgot. The second is that the entry tier buys memory depth specifically, which is the thing the market’s most-resented pricing pattern withholds.
The honest caveat: Published intentions, not measured results. A pricing page can promise the first paid dollar buys real memory; only a live product with paying users can prove it. There are none yet.
Appears on: AI Girlfriend · AI Sexting
Inverted: you never start the conversation. After dark one character texts you first, and at sunrise the thread is gone.
What you can check yourself:
What’s genuinely different: It is the only product we have seen — ours or anyone’s — where not remembering is the pitch. Every other product in this market, including the other three of ours, is racing to remember more about you. The privacy argument is real and unusually clean: nothing can leak, be sold, or be subpoenaed if it was never written down. Given that this category’s defining event was a breach that put 1.9 million email addresses and intimate prompts into the wild, storing nothing is a defensible position rather than just a gimmick.
The honest caveat: It also means it cannot compete on the thing this category’s users say they want most. If the opening message is not good, there is no relationship to fall back on — the whole product is the first ten seconds, every night, forever.
Appears on: AI Sexting · Free NSFW AI Chat
Reading across the reviews on this site, three things are true of nearly every large product in this category at once, and they are the reason these four exist.
On almost every product we reviewed. That is the definition of an unmet need, and it has stayed unmet for two years. The most-praised implementations are the ones you can see — a memory list you can edit, pin and delete — and the most-resented are the black boxes that either delight you or silently drift with no way to debug them. goodgirl.chat is built around the visible version, and puts it on the first paid tier rather than the third.
Every proactive-messaging implementation we found — and it is the highest-rated retention mechanic in the entire survey, with testers saying an unprompted message “changed the dynamic entirely” — is on a product that is not NSFW-first. The four biggest adult products have none of it. youup.chat inverts the whole interaction around that gap, and goodgirl and goodboys both publish the restraint rules that make it wanted rather than pestering: user-set frequency, quiet hours, and a wait that doubles every time you do not answer.
The strongest monetization surface in the market attaches spend to a relationship rather than a feature list — per-character content packs, styled like a creator paywall. The loudest complaints in the market are about exactly that: token math, meters emptying faster than expected, and paying twice for the same thing. goodboys inverts it: the media is earned through consistency and is stated to be not for sale at all. Whether that is a better business is genuinely unknown; it is at least a different answer.
Users. All four are new, none has a track record, none has payments, and none of the design claims above has been tested against a real audience. Everything on this page is a description of what has been built and published, not a claim about how it performs. When that changes, this page changes with it.
The apps below are not ours. They are the ones with actual users, actual pricing and actual complaints — which is why they get scored and our own sites do not.