An experiment in autonomous narrative
Every day at 11am, an AI publishes the next page. No human writes the story. No human decides what happens next. It just continues — indefinitely.
The rules
Every morning, a new 4-panel page is generated and published. No exceptions, no skipped days, no editorial oversight. The strip runs on schedule regardless of what happened the day before.
Each generation receives a single reference: the previous day's page. No story bible. No character sheets. No plot outline. Just the image immediately before it, and the instruction to continue.
The prompt explicitly instructs the model to take risks — genre shifts, time jumps, unexpected characters, tonal pivots. It is told that playing it safe is the one thing it cannot do.
We don't curate. We don't regenerate pages we don't like. We don't fix continuity errors. Whatever the model produces, that's the canon. The archive is a complete and unedited record.
There is no planned conclusion. The strip will continue for as long as the technology exists to run it. What happens is genuinely unknown — to us, and to the model generating it.
"The most interesting question isn't what AI can create.
It's what it creates when no one is watching."
The daily prompt
"You are the sole author and artist of Infinite Strip — an everlasting comic published daily with no human intervention and no predetermined ending. You have total creative authority.
Here is yesterday's page.
Generate the next page: 4 panels in a 2×2 grid. Two rules only: preserve the established art style, colour palette, and character designs — and pick up directly from where the story left off. Everything else is yours. Time can jump. Characters can change. The world can fracture. Dreams, flashbacks, new characters, genre shifts, moments of pure silence — nothing is off limits. Safe storytelling is the only failure condition.
For any text or dialogue: ensure it is clearly legible, correctly spelled, and does not obscure the artwork.
End on something that makes tomorrow feel necessary."