이미지: GitHub — 신규 인기 AI 레포
Summary
- GitHub user 0toshigami has open-sourced 'orbis-pictus,' an interactive picture book built on real-time AI image generation
- Type text or click anywhere on the image, and AI instantly draws a new page reflecting that input
- The repository states it is an unofficial homage to a service called flipbook.page and is not affiliated with that team
- 저장소명
- 0toshigami/orbis-pictus
- 공개 방식
- 오픈소스, MIT 라이선스
- 확보 스타 수
- 26개(포크 13개)
- 핵심 동작
- 텍스트 입력·이미지 클릭 시 AI가 새 페이지 이미지를 실시간 생성
- 오마주 대상
- flipbook.page (공식 관계 없음)
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 shown on screen isn't HTML — it's a single AI-generated image. When a user types any sentence, a picture page depicting that content is created, and clicking anywhere on the image generates a new page centered on that spot. There are no links, no markup. Every pixel is redrawn each time a request comes in, the description says. The repository has received 26 stars and 13 forks and is distributed under the MIT license. The repository's documentation states that the project is an independent, unofficial homage to a service called 'flipbook.page' and is neither created nor endorsed by that team.
What this means
Until now, "browsing" the web has meant following pages and links someone else built in advance. orbis-pictus flips that premise. There are no pre-drawn pages; an image matching the spot you click is generated only at the moment of interaction. 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 continuation scene with every interaction. What makes this hard is a combination of speed and consistency. Users can only tolerate a short wait, yet the image also needs to flow naturally from the previous page for it to feel like a storybook. That's likely why the picture-book metaphor was chosen — each page turn (or each poke into the image) branches the story further. The backdrop here is that as image generation models have gotten faster in response time, it's becoming feasible even for 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 might be consumed. Instead of flipping through fixed pages, an experience where pressing on whatever piques your curiosity generates a story on the spot could extend 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's worth watching whether this kind of click-driven generative interface spreads to other creative tools.
