Anvaya

Inspiration

Food is the most intimate archive of culture, and it's disappearing. Recipes passed down through generations were never written down. They lived in grandmothers' hands, in the smell of a kitchen, in a pinch of something unnamed. When those people are gone, the recipes go with them. We built Anvaya because we believe that memory is enough to start, and community is enough to finish. The name means lineage in Sanskrit, the thread that connects you back.

What it does

Anvaya is a cultural recipe memory recovery app. You describe a dish someone made for you like the smell, the texture, who made it, where you were, etc., and Anvaya uses AI to reconstruct it step by step. The jigsaw mechanic keeps each step locked until you complete the one before it, so you discover the recipe as you cook it. Nani, our AI cooking agent, guides you through voice and a live video call that uses computer vision to scan your dish in real time and tell you what to fix. A community feed, where failures are celebrated as loudly as wins, keeps you accountable and connected. Your Journey tracks how close your version is to the original, your Library holds every recipe you've saved or recovered, and your Profile builds a living taste map of who you are as a cook.

How we built it

We designed Anvaya in Figma, building a full mobile design system rooted in warmth, Playfair Display for the emotion, Inter for clarity, and a cream-and-terracotta palette inspired by handwritten recipe cards and wooden spoons. Every interaction was designed to feel like opening a letter, not using an app. We used Claude AI to prototype the Nani conversation flows and AI scanning logic, and mapped the full user journey from first memory all the way to sharing a dish with the community.

Challenges we ran into

The hardest design challenge was the failure-positive community, because most social apps reward perfection. We had to intentionally architect a system where "beautiful disaster" carries the same visual weight as "nailed it," without it feeling performative. The second challenge was making the Nani video call feel like a FaceTime with a real person, not a chatbot interface. Getting the tone and intimacy right, warm but intelligent, took significant iteration.

Accomplishments that we're proud of

The 68% flavor match progress ring on the Journey screen, which is a completely novel way to visualize how close your version is to a memory
The jigsaw recipe mechanic recipe steps that unlock only as you cook, turning a recipe into an experience
Designing a failure-first community where disasters are featured, not hidden
The Nani AI video call with real-time dish scanning and ambient analysis cards
A design system warm enough to feel like home, sharp enough to win a designathon

What we learned

Cultural food is deeply emotional, and that emotion has to live in the design, not just the copy. We learned that the most important design decision we made was slowing things down: one question at a time, one step at a time, one memory at a time. We also learned that diversity in a product like this isn't an afterthought. Jollof Rice, Mole Negro, Dashi, and Dal Makhani all deserve equal reverence in the same feed.

What's next for Anvaya

[ ] Voice-first memory intake, where you speak your memories instead of typing them
[ ] Real ML model trained on regional cuisine databases to improve recipe reconstruction accuracy
[ ] Nani video call with live speech-to-instruction AI (not just visual scanning)
[ ] Collaborative recovery, where you invite a family member to add memories to the same dish
[ ] Regional dialect support for Nani's voice (Tamil, Yoruba, Spanish, Japanese)
[ ] "Recipe will" feature where you can record a dish as a legacy for someone you love
[ ] Partner with cultural preservation organizations and diaspora communities