I Never Met My Naanu. So I Built a Place to Remember Him.🌼

A developer built MemoryDay · Our Ofrenda, an open-source, Día de los Muertos-inspired digital remembrance altar and AI-assisted family keepsake book generator that runs Google's Gemma model locally via Ollama. The app stores family photographs, stories, recipes and voice keepsakes in browser-based IndexedDB under a local-first design, and is explicitly built so the AI acknowledges missing information rather than inventing memories. It is deployed on Render with source available on GitHub.

An altar built with code. A family archive built with love. A little piece of forever. Some people leave this world, but they never really leave our lives. They remain in the recipes we still cook, the songs that suddenly bring tears to our eyes, the stories we repeat at family gatherings, and the photographs we hold a little longer than we should. But what happens when the people who remember those stories are no longer here to tell them? What happens when a grandmother's recipe is forgotten, a grandfather's voice disappears, or a beautiful family memory gets lost among thousands of photographs? That question became the heart of MemoryDay · Our Ofrenda . I didn't want to build just another AI application. I wanted to build something that could preserve the little things that make someone unforgettable. Something that could turn memories into something we can revisit, share, and pass down. And that's how MemoryDay was born. MemoryDay · Our Ofrenda is a digital Día de los Muertos-inspired remembrance altar and an AI-assisted family heirloom book generator. It brings together cultural remembrance, family storytelling, personal archives, and open-source AI to create a space where memories can live on. The idea is simple: Technology should not only help us remember information. It should help us preserve the people behind those memories. With MemoryDay, families can create a personal digital ofrenda for someone they love. The keepsake book is especially close to my heart. It transforms photographs, stories, recipes, and personal memories into a lasting family document that can be printed or preserved digitally. Because sometimes, a digital file isn't enough. Sometimes, you want something you can hold. Explore MemoryDay here: Live Application: https://memoryday.onrender.com https://memoryday.onrender.com The experience is best explored by creating a remembrance altar, adding memories, and exploring the keepsake book. GitHub Repository: https://github.com/AaliyaKhan25/MemoryDay https://github.com/AaliyaKhan25/MemoryDay The project is open source. You can explore the implementation, understand how the features work, and contribute to making it better. I wanted MemoryDay to feel personal, peaceful, and respectful. Every design and technical decision was made with that experience in mind. I designed the interface around the visual language of a traditional ofrenda. Deep plum tones, golden accents, candlelight, marigolds, and warm details help create an atmosphere of remembrance rather than making the application feel like an ordinary dashboard. The goal was to make the experience feel like entering a meaningful space. One of the most meaningful parts of this project is its use of Gemma , Google's open-weight model family. Through a local Ollama setup, Gemma powers AI-assisted experiences such as working with family memories and generating content for the memorial keepsake experience. I wanted to explore how open-weight AI could help people interact with their own family archives without making private memories dependent on a mandatory cloud AI service. The Family Archive is designed around a particularly important principle: AI should not invent someone's memories. If information is not present in the available family archive, the experience should acknowledge that rather than manufacture an answer. Family photographs, personal stories, and voice keepsakes are deeply personal. They deserve to be treated with care. MemoryDay uses browser-based IndexedDB for local storage of personal content and is designed around a local-first approach. The intention is to keep private family memories under the user's control rather than requiring them to upload everything to a remote service. Local AI functionality depends on having the required local model and Ollama setup available. MemoryDay is deployed using Render , making the application accessible through a live web experience. Live Demo: https://memoryday.onrender.com https://memoryday.onrender.com I believe open innovation matters because meaningful technology should not be limited to people who have access to expensive tools, private infrastructure, or large development teams. Open-source development gives people the opportunity to learn from one another, improve existing ideas, and build things that can make a real difference. For a project like MemoryDay, openness means more than sharing code. It means allowing developers to explore how AI can be used for cultural preservation, personal storytelling, and family history. It means encouraging conversations about privacy, responsible AI, and the importance of keeping human experiences at the centre of technology. And most importantly, it means that an idea created by one person can become something that many people help improve. I am proud that MemoryDay is part of that open-source journey. Building MemoryDay involved exploring how AI could become a meaningful part of a personal application rather than simply being added as a feature for the sake of using AI. I worked on bringing together the user experience, local-first architecture, family archive, and Gemma-powered experiences into one application. The most important design consideration was ensuring that AI supports family memories instead of replacing or inventing them. There is something deeply human about remembering. We remember people through the food they made, the advice they gave, the songs they loved, and the little habits that once seemed ordinary. A photograph can preserve a face, but it cannot always preserve the story behind it. A voice recording can preserve a voice, but it cannot explain why that voice meant so much. A recipe can preserve ingredients, but it cannot always preserve the person who taught us how to make it. I wanted MemoryDay to bring these little pieces together. I wanted someone to be able to open the application years from now and find more than a photograph. I wanted them to find stories, letters, recipes, voices, and memories that make a loved one feel close again. This project also made me think differently about AI. We often talk about AI in terms of productivity, automation, speed, and intelligence. But what if we also used it to protect the things that make us human? What if technology helped a future generation hear the stories of a generation they never got to meet? What if it helped a family preserve a recipe that would otherwise disappear? What if it gave memories a place to stay? That is what I hope MemoryDay can become. It is not about replacing remembrance. It is about giving remembrance another place to live. Because people may leave our sight, but the stories they leave behind deserve to stay. I am submitting MemoryDay for the following categories: MemoryDay started with a simple thought: memories should not disappear just because time passes. I built this project with the hope that it could help families preserve the stories, traditions, photographs, and voices that make their loved ones unforgettable. It is still a work in progress, and there is so much more I would love to add. But every project has a reason for existing, and mine is this: Some applications help us manage our lives. Some help us build our future. I wanted to build one that helps us hold on to the people who made us who we are. Thank you for taking the time to explore MemoryDay. 🌼 Built with love by Aaliya Khan. MemoryDay · Our Ofrenda — A digital altar. A family archive. A legacy of love.