Inspiration
Anorexia nervosa has the highest mortality rate of any psychiatric illness, yet the tools built for recovery almost entirely ignore a core mechanism of the disease: broken interoception. Research shows that people with eating disorders have an average interoceptive deficit effect size of 1.62 meaning they genuinely cannot feel what their body is communicating. We were inspired by the question: what if instead of asking patients to trust a perception system that's been damaged by the illness itself, we built something that could speak for the body in a language it couldn't distort?
What it does
Attune translates eight real-time biosignals glucose, ghrelin, leptin, gut motility, cortisol, HRV, body temperature, and long-arc leptin into a living 3D garden of flowers. Each flower represents a signal: a Poppy for ghrelin, a Sunflower for temperature, a Lotus for long-arc leptin. When a signal moves outside its contextual desired state, petals fall. When it stabilises, the garden blooms. Patients also receive short, felt-language body voice messages ("I'm rising. That restless feeling has a name.") that witness rather than instruct. Attune is prescribed alongside clinical treatment not instead of it.
How we built it
We built Attune through a tight loop between three AI tools Claude, ChatGPT, and Figma Make using each and moving components fluidly between them.
Challenges we ran into
The hardest design challenge was building something that felt safe and non-clinical for a population that is hypervigilant about body data. Every design decision the felt language, the petal metaphor, the absence of numbers was in service of reducing threat response while still conveying real signal. Technically, getting the flower card grid, animations, and drop zone to behave consistently across iterations required very precise, incremental prompting.
Accomplishments that we're proud of
We're proud of translating a genuinely complex neuroscientific and clinical concept into something that feels warm, intuitive, and beautiful. The body voice copy for Scenario 1 five signals, five lines, each one a witness not an instruction is something we're particularly proud of. We're also proud of grounding every design decision in peer-reviewed science while keeping the experience human-first.
What we learned
We learned that the gap between what data shows and what a patient perceives is not just a clinical problem it's a design problem. We also learned that restraint is the hardest design skill: every time we were tempted to add more information, more numbers, more feedback, the right answer was almost always less. The body voice tone witness, not instruct became a guiding principle that shaped the whole product.
What's next for Attune
Honestly, we had a lot of fun building this, and that's where we want to start. The next step is figuring out whether Attune can move from speculative concept to something real: could the biosignals actually be captured? Could the flower metaphor work in a clinical setting? Could this live in a different context entirely chronic illness, anxiety, pain management anywhere interoception breaks down? We'd love to put it in front of clinicians and patients and find out. For now, we're proud of what the concept asks: what if recovery tools were designed around perception, not just behaviour?
Built With
- chatgpt
- claude
- figma
Log in or sign up for Devpost to join the conversation.