Diffuji
a diffusion-powered instant camera
Last verified:
What is Diffuji?
Diffuji is an instant camera concept built at TreeHacks 2026 that combines a Raspberry Pi Zero 2 W, a camera, and a thermal receipt printer inside a custom housing. A user snaps a photo, chooses a mode, and the image is sent to an AI backend that transforms it before printing the result on receipt paper.
The core idea is to make photography playful and surprising. The website highlights filter modes that use diffusion models to restyle photos, including stylized looks like studio Ghibli-inspired rendering, time-travel effects, duck transformations, exaggerated muscles, and adding Sam Altman into the image.
Diffuji also includes search modes that use Perplexity web search to analyze what the camera sees. Those modes can estimate the price of visible items or identify objects and landmarks, turning the camera into a visual question-answering device as well as a creative photo printer.
The project is positioned as a fun, experimental demo rather than a conventional consumer product. It was built as a hackathon project and won both the Neo Prize and the Most Creative Prize, so it is best suited to people interested in AI image generation, playful hardware prototypes, and interactive demos.
Diffuji pricing
Pricing model: Freemium
The website does not list any pricing, free tier, subscriptions, or paid plans. It presents Diffuji as a hackathon-built prototype and demo, with no purchase or usage pricing disclosed. The site also does not describe included plan features, credits, or limits, so there is no official commercial pricing structure shown.
Diffuji pros
- Instant photo-to-print workflow
- Combines camera and thermal printer
- Custom handheld hardware build
- Supports creative image restyling
- Uses diffusion models for transformations
- Offers multiple visual modes
- Includes studio Ghibli-style effects
- Can time-travel photos to past decades
- Can turn people into ducks
- Can exaggerate muscles in photos
- Can insert a specific public figure into an image
- Includes search-based analysis modes
- Can estimate visible item prices
- Can identify objects and landmarks
- Lets users A/B test prompts across four AI providers
- Uses OpenAI image generation backend
- Uses Gemini image backend
- Uses Modal Flux Kontext backend
- Uses Perplexity search backend
- Prints outputs on receipt paper
- Inkless printing via thermal printer
Diffuji cons
- Hackathon prototype, not a finished consumer product
- Limited to the modes shown on the site
- Output quality depends on backend AI behavior
- Thermal prints are low-fidelity compared with photo prints
- Receipt paper format is physically small
- Requires custom hardware to operate
- Depends on AI backend availability
- Search modes rely on visible scene content
- Not designed for general-purpose photography
- Some transformations are novelty-focused rather than practical
- No pricing or commercial availability listed
- No app store or standalone software product described
- No privacy policy or data handling details shown on the site
- No battery-life or portability specs are provided
- No support documentation or user guide is shown
Frequently asked questions about Diffuji
What is Diffuji?
Diffuji is an instant camera project that captures a photo, sends it to an AI backend, transforms it in a selected mode, and prints the result on thermal receipt paper. The site describes it as a diffusion-powered instant camera built at TreeHacks 2026.
How does Diffuji work?
You take a picture, choose a mode, and the image is routed to an AI backend for processing. After the transformation, the camera prints the output on receipt paper through the built-in thermal printer.
What kinds of image modes does it have?
The site shows filter modes that restyle photos with diffusion models, including studio Ghibli-style rendering, time-travel looks, duck transformations, muscle exaggeration, and adding Sam Altman into the photo. It also mentions many more similar creative modes.
Does Diffuji analyze the real world too?
Yes. The site lists search modes that use web search to analyze what the camera sees, including estimating the price of visible items and identifying objects and landmarks.
Which AI providers does Diffuji use?
The site says each request can be routed to one of four AI providers for A/B testing: OpenAI, Gemini, Modal, and Perplexity. This lets the same prompt be compared across different backends.
What hardware is inside Diffuji?
The website says it uses a Raspberry Pi Zero 2 W, a camera, and a thermal receipt printer inside a custom housing. It is described as a compact hardware prototype rather than a purely software tool.
What does Diffuji print on?
Diffuji prints the transformed image on receipt paper using a thermal printer. The site highlights that this is inkless printing.
Is Diffuji a commercial product?
The website presents Diffuji as a hackathon project built at TreeHacks 2026, not as a packaged commercial product. It does not list storefront availability, licensing, or consumer pricing.
What makes Diffuji different from a normal instant camera?
A normal instant camera prints what it captures, while Diffuji sends the photo through AI first. That means the print can be creatively reimagined or analyzed instead of being a direct capture.
Why was Diffuji made?
The site frames the project as a playful experiment in making instant cameras more imaginative and surreal. It emphasizes creativity, novelty, and AI-powered transformation of everyday moments.