Swaplse is an AI generator for adult images and video: you describe a scene and get it back in about a minute. 4,580 works are published here right now - 2,324 clips and 2,265 stills - so you can judge the output before writing anything. The first generation of each mode is free.
The search tail around SpicyChat is almost entirely login attempts and the Android build. If you are trying to get back into an account, nothing here does that. If you are looking for something else to use, the difference is worth two sentences.
Browser, not an install
Swaplse is a website. There is no APK to find, no version to keep updated and no permissions dialog. It behaves the same on a phone and on a desktop, and the library plays video inline while you scroll rather than making you download first.
What it does instead of chatting
You write a scene, it renders it. Three modes: text to image, text to video, and photo to video for a frame you already have. Video is a first-class mode - 2,324 of the 4,580 published works are clips.
Writing the first prompt
Four clauses do most of the work: who (one subject, plainly described), what is happening (one action, present tense), where (a room, a surface, a time of day), and how it is shot (close-up or full body, soft or hard light). The last clause is the one people skip and the one that decides whether the result looks photographed or rendered.
Coming from a chat product
The habit to unlearn is length. In chat, more context is better; here a paragraph of backstory is ignored and a crisp frame description wins. If it is the conversation you actually want, that is girli.app - a separate site of ours built for it.
Cost
First generation of each mode free, no card. Browsing the library free with no account. After that it runs on coins you spend deliberately.
Start
Pick image or video, type the scene, wait about a minute. The first one of each mode is free, so finding out whether it suits you costs a minute.
A chat scene has no camera in it
The thing missing from every log you might want to turn into a picture is the viewpoint. Prose leaves it to the reader, and an image model will not choose for you - it defaults to a flat eye-level shot at conversational distance, which is why converted chat scenes come back looking oddly polite. Two clauses fix it, and they are the two the conversation never needed: where the camera is, and where the light is.
Picking the moment, not the scene
A scene runs for paragraphs; a frame is one instant. Choosing that instant is the whole job, and the instant before or after an action almost always renders better than the action itself, because a held pose has a shape the model can resolve.
- Take one sentence from the log, not a summary of the exchange.
- Add the viewpoint:
low angle from the foot of the bedbeats any adjective about intensity. - Add one light source and its direction:
single warm lamp behind her, left side.
The mistake carried over from a chat log
Dialogue lines left in the prompt are the common mistake. Quoted speech is not ignored - it is treated as a description, and the model sometimes tries to render the words themselves as text on a surface in the frame. Lettering is the one thing both engines are reliably terrible at, so any run that comes back with garbled writing in it usually has a quotation mark still in the prompt.