This page is the long version of what happens after you press Generate. There are two models rather than one, they run in sequence, and most of what makes the output usable happens in the first pass — before anything is drawn. It takes about two minutes end to end.
One to six photos, an optional story template, an art style, a mood, and a line of context if you want one. Everything after this is derived from those inputs; nothing is picked at random between runs.

Before any image exists, a language model produces a script: the number of panels, what happens in each, who speaks and what they say, and a character anchor — a fixed written description of each person in your photos.
Every panel is rendered against that same script and the same character anchor, with a shared seed and one mood and setting for the page, so the person in panel one is the person in panel three.

Your browser is sent a blurred, watermarked copy. The full-resolution file stays in private storage until a payment is confirmed, then it is released to you.
Real inputs and real output. The likeness carries over; the pose and background do not, because the result is drawn rather than filtered.




The same photo produces a different comic in each. Switching style costs nothing before you pay.
Dreamy anime
Manga
Heroic Comic
3D Animated
Classic comicAsking one image model for four panels gives you four unrelated pictures: different faces, no sequence, dialogue that does not follow. Separating the work fixes the cast, the order and the words before a single pixel is drawn, and gives every panel the same thing to draw against. It costs one extra model call per comic and it is the difference between a page and a pile of images.
It is a paragraph of plain text describing each person — face, hair, build, clothing — written once by the writer model and passed verbatim into every panel request. Panels are never re-derived from your photo individually, because doing that produces a slightly different person each time. This single detail is responsible for most of the consistency you see across a page.
A watermark over a full-resolution image can be cropped or painted out, and a blur applied by the browser can be switched off in one click. So the blur is applied to the pixels on the server, before they are sent. The sharp version genuinely is not in your browser until you pay, which is why the preview costs you nothing and still protects the work.
A page that fails to draw is retried, and one failed page does not discard the pages that worked. If you pay and your browser closes before the download finishes, the comic is still on the server for 24 hours and is re-delivered from your receipt the next time you open the site. After 24 hours it is deleted whether or not it was bought.
Under two minutes in the usual case, from uploading the photo to a finished preview page on screen. Extra pages are drawn after payment and take roughly the same each. Nothing is queued for later and nothing is emailed to you — you wait on the page and it appears.
No. Generating, previewing, paying and downloading all work without an account. The receipt is held in your browser rather than against a login, which is also why there is nothing to cancel later and no password to reset. Download the file on the device you paid on.
Not on the same comic — starting over generates a fresh one. The preview is free, so trying a different style or context costs nothing but the time.
They are passed to the generation model in the request and never written to disk on our side. The finished comic is held for 24 hours so it can be delivered, then deleted.
Last updated . Prices and limits on this page are checked when it changes.
Finished output, not mock-ups.



Generating and previewing costs nothing and needs no account. Downloads start at $3.99, one time.
We use analytics cookies to see which pages work and which do not. Nothing is set unless you agree, and we never sell your data. Privacy Policy