Open-source picture book where AI draws a new page with every click

What was released

GitHub user 0toshigami has released an open-source project called 'orbis-pictus.' According to the repository description, the app calls itself an "infinite browser." Every page on screen is not HTML but a single AI-generated image. When a user types any sentence, a picture page reflecting that content is created; clicking anywhere within the image triggers a new page to be drawn for that spot. There are no links and no markup — every pixel is described as being redrawn on each request. The repository has 26 stars and 13 forks and is distributed under the MIT license. The repository documentation states that the project is an independent, unofficial homage to a service called 'flipbook.page' and is not created or endorsed by that team.

What this means

Until now, "browsing" the web has meant following pre-built pages and links that someone else created in advance. orbis-pictus flips that premise. There are no pre-rendered pages; an image fitting the spot is generated only at the moment of the click. Technically, this appears to combine an image generation model not just with text prompts but with visual coordinate information — the click location — to instantly produce a plausible follow-up scene for each interaction. This is difficult for two reasons: speed and consistency. Users can only tolerate short wait times, yet the image also needs to flow naturally from the previous page to feel like a storybook. That's likely why the picture-book metaphor was chosen — each page turn (or poke into the image) branches the story further. The recent speedup in image generation models' response times can also be seen as part of the backdrop, making it feasible for even individual developers to attempt this kind of "draw-as-you-explore" interface.

Why it matters

orbis-pictus itself is a hobby-scale homage project, but the idea behind it can be read as an experiment in how content is consumed. Instead of flipping through fixed pages, an experience where clicking on a point of curiosity generates a story on the spot could have applications ranging from children's picture books to interactive fiction. Since it's open source, anyone can grab the code and build their own "infinite picture book," and it will be worth watching whether this kind of click-driven generative interface spreads to other creative tools.