Memory Hive | Devpost

Inspiration

Shortly before LA Hacks, we met each other for the first time, and wanted to connect and bond so we could form a stronger team. However, we found it difficult to connect beyond the realm of small talk in the short amount of time we had before the hackathon, and even though two of us knew each other beforehand, we discovered we didn’t know much about each other. Thus, we asked ourselves the question: how can we facilitate connection and build meaningful relationships with the people around us? The rise of social media has sparked a loneliness epidemic, so we wanted to allow people to share memories and stories through the medium of meaningful objects in order to facilitate deeper human connection and community building, and bring memories to life.

What it does

We created a Snap AR Lens that allows users to design a personalized virtual room where they can place cherished objects connected to meaningful memories. Each object represents a special story or moment, helping users build a unique space filled with personal significance. By blending augmented reality with memory-sharing, we aimed to create an immersive and emotional experience where every room becomes a living archive of the user’s most treasured connections.

How we built it

We built Memory Hive by extending Snapchat’s Remote Object Detection project sample, which provided a strong foundation for the features we envisioned. Using Lens Studio, we wrote custom TypeScript scripts to add functionality, including interactive buttons for user actions. For object detection and management, we integrated Hugging Face models, and to populate the memory rooms, we used Meshy AI to generate 3D models of meaningful objects. Together, these tools allowed us to create a personalized, interactive AR experience centered around memories and emotional connection.

Challenges we ran into

Using Lens Studio had a steep learning curve for us. The interface was unintuitive, the software frequently crashed, and we initially installed an incompatible version, forcing us to uninstall and reinstall to even get started. Additionally, much of our team didn’t have a lot of AR experience, so we spent a significant amount of time learning the basics and figuring out how to build the features we wanted. Lastly, working with many large 3D models introduced us to GitHub Large File Storage (LFS) for the first time, which added another layer of complexity to our development process.

Accomplishments that we're proud of

Although we struggled to implement our idea, we’re proud of having a strong concept to begin with. We formulated a unique and meaningful idea that we all felt strongly about: we all believed in the power of human connection through shared experiences and storytelling, and bringing memories to life. We’re also proud of our resilience and collaboration as a team: when it would have been easy for us to get frustrated and blame each other for our shortcomings, we instead collaborated, persevered, and came together as a team to tackle challenges head on.

What we learned

Over the course of the project, we gained hands-on experience with Lens Studio, where we designed and built interactive functionality for Snap AR Lenses. In addition to developing creative AR experiences, we also learned how to implement TypeScript scripting within Lens Studio to add more complex behaviors and custom interactions. To manage our project files more effectively, especially larger assets like 3D models and textures, we became proficient in using GitHub Large File Storage (LFS), ensuring smooth version control and collaboration across the team.

What's next for Memory Hive

For this project, we were only able to use pre-generated 3D models as assets because Lens Studio doesn’t support adding assets at runtime. In the future, we hope that Snapchat can add functionality to allow users to scan objects from the real world and add these assets at runtime. We would also add multiplayer functionality so that users can visit each other’s rooms and multiple users can view the same room at a time, creating a true “hive” and allowing for community formation. Lastly, we would connect to Gemini API to provide AI descriptions of the historical, social, or cultural significance of the items in the room to provide more context and open up deeper opportunities for meaningful connection.