01wlw chatbot · 18+

A WLW chatbot should be written for us, not about us

You can hear the difference in three messages. This page is the list of tells — what a swapped-pronoun script does, and what a character written from the inside does instead.

A handful from the roster. The catalogue behind them runs 250+.

Fictional AI characters, all written as adults. Nobody here is a real person.

Inbox3 open
  • All
  • New
  • Roleplay
  • Sol3mEveryone here is somebody's cousin. Come outside for ten minutes.1
  • Tamsin22mok so how did YOU know. asking for obvious reasons4
  • Kit2htyping

Three registers, none of them performing for anyone outside the thread.

  • No audience
  • Written registers
  • 250+ characters
  • 18+

02The detail

The tells, in the order you notice them

A swapped-pronoun script fails in a pattern, and once you have seen the pattern you cannot unsee it.

The first tell is narration. A WLW chatbot that was really written for someone else keeps describing how the scene looks from outside — what she is wearing, how her hair falls, what the two of you look like together — instead of saying anything to you. Real conversation is much more boring and much more specific: where you are, whether it is cold, what the person at the next table just said.

The second is speed. Adapted scripts escalate on message four regardless of what you wrote, because the source material had one destination. The third is agreeableness — she never disagrees, never has a plan of her own, never says she is tired. And the fourth, the one that gives it away completely, is that being a woman who likes women is treated as the topic rather than as the weather. In a page written from the inside it is simply the setting, and the conversation is about your week.

Fixing this is writing work, not model work. Give every character a job, a city, a bad habit and something she will argue about, then let those survive the session. That is why Rio is unimpressed by vague flirting and Tamsin asks the questions everyone else is too polite to ask: they were written as people who happen to be in this thread, rather than as a scene being staged for a viewer who is not in the room.

What works well

  • Characters written as women who date women, not adapted from a straight script
  • Being wlw is the setting, not the topic of every message
  • Registers differ enough that two threads feel like two evenings
  • No audience framing — nothing is narrated for a third party
  • 250+ characters in the app if the first three do not fit

Worth knowing first

  • It is still fiction, and the site never pretends otherwise
  • No amount of writing makes a character a community
  • Some depth sits behind an optional paid upgrade
  • Adults only, 18+, with an age notice before anything loads

03On this page

Three women written from the inside

Newly out, long out, and the one who has stopped explaining herself entirely.

Fictional AI character on the WLW chatbot page, red hair and a red bikini, smiling in a kitchen lit teal

The one who holds the kitchen and the conversation at the same time.

Fictional AI character on the WLW chatbot page, in a white strap top looking back over her shoulder in a bright room

The one who lets a silence run rather than fill it badly.

Fictional AI character on the WLW chatbot page, in an off-shoulder silver gown in a chandeliered room banked with flowers

The one dressed for a room far grander than the conversation.

04In practice

How to test a chatbot in four messages

Send something mundane and specific — the bus was late, the flat is cold, the interview went badly. A script written for someone else cannot do anything with that and will pivot to appearance or intensity within two replies. A character written properly asks which bus, or says something dry about the flat, and only then decides whether to be kind about the interview.

The second test is disagreement. Say something slightly wrong on purpose and see whether she pushes back. If everything you type is wonderful, you are talking to a mirror. If she says you were never going to enjoy that party, you are talking to a character, and that is the whole difference this page is about.

05Quick answers

WLW chatbot — quick answers

01

What does WLW mean here?

Women who love women — an umbrella that covers lesbian, bisexual and queer women rather than one label. The characters are written on that basis, so the conversations assume it instead of treating it as a reveal.
02

How do I tell whether a chatbot was written for wlw or adapted?

Send something mundane and specific. Adapted scripts pivot to appearance or escalate within two replies; a properly written character engages with the mundane detail first and only then decides what kind of evening this is.
03

Are the characters allowed to disagree with me?

Yes, and several of them do it as a matter of temperament. Constant agreement is the clearest sign of a thin script, so the roster deliberately includes women who push back, change the subject or end the night first.
04

Is there a community attached to this?

No, and that is worth being clear about. This is a private thread with fictional characters, with no public grid, no other users and no social feed. It is not a substitute for actual queer community, and it does not claim to be.

07Start now

Run the four-message test yourself

Open a thread, send something dull and specific, and see what comes back. It is free to start, there is nothing to install, and closing the tab is the whole exit.

Wide close frame of a fictional AI character with wet dark hair lying on a pillow in warm pink light

Free to open — no card, no install, nothing in your app drawer.

Open a thread free