Inspiration
When we lose someone we love, we don't only lose their photos or messages. We lose the small conversations too asking for advice, telling them about our day, or simply saying "I miss you."
We thought: what if we could still talk to them through the memories they left behind?
That is how we came up with Still With You.
What it does
Still With You lets people preserve the memories, stories, advice, values, and everyday moments of someone they love.
Later, an authorized family member can talk to that person's AI memory space, ask questions, revisit moments, or simply have a conversation.
The goal isn't to replace the person. It's to keep a connection possible through what they left behind.
How we built it
We use CockroachDB as the persistent memory layer. Memories are stored with vector embeddings, and Distributed Vector Indexing helps find memories that are relevant to a question.
The retrieved memories are then used by the AI to generate a response grounded in those memories.
We also used AWS Lambda for the backend and AWS Amplify to host the application. During development, we connected Cline to our CockroachDB cluster through the Managed MCP Server to inspect and verify our database.
Challenges we ran into
The main challenge was making the AI feel connected to a specific person instead of behaving like a normal chatbot.
We had to make sure the right memories were retrieved before generating a response, while also keeping different family members' access separate.
Accomplishments that we're proud of
We're proud that we built a working conversational memory system where information can persist beyond a single chat session.
A question doesn't need to use the exact words from a stored memory. The vector search can find related memories and bring them into the conversation.
We're also proud of connecting CockroachDB's relational and vector capabilities with AWS and an AI model into one working application.
What we learned
We learned that an AI conversation is only as meaningful as the memories behind it.
Building Still With You also showed us how important persistent memory, semantic search, and access control are when creating personal AI experiences.
What's next for Still With You
We want to make the experience richer while keeping it personal and safe including more ways to preserve memories and eventually making conversations feel even more natural.
The idea is simple: the person may be gone, but the conversations and memories they left behind can still mean something.
Built With
- amazon-web-services
- amplify
- cockroachdb
- css
- gemini
- generative
- groq
- html
- javascript
- lambda
- mcp
- node.js
- search
- vector
Log in or sign up for Devpost to join the conversation.