• 

• 

• 

• 

• 

Inspiration

In meetings and lectures, important moments often pass by because we’re focused on listening rather than taking notes. Existing tools either require constant manual input or record everything, which can be distracting and raise privacy concerns. We wanted a system that quietly remembers what matters without interrupting the moment.

That idea became NovaMind.

What it does

NovaMind listens for important moments and turns them into sticky notes you never forget.

It passively detects:

Action items

Key insights

Important reminders

Instead of recording full conversations, NovaMind selectively captures only meaningful moments and externalizes them as short, actionable sticky notes that persist using memory.

How we built it

NovaMind is a full-stack application built with:

Next.js + React + TypeScript for the frontend and backend

Supermemory for persistent memory storage

LLM-based intent extraction to detect important moments

Web Speech API for privacy-first speech-to-text

Tailwind CSS for a clean, minimal UI

Audio is processed locally in the browser, and only short, intent-relevant text snippets are sent to the backend—no raw audio is stored.

Challenges we ran into

Designing selective capture without being intrusive

Maintaining strong privacy guarantees

Structuring memories so they are useful and retrievable

Integrating external SDKs while keeping strong type safety

Accomplishments that we're proud of

Building a complete full-stack memory system in a short time

Achieving privacy-first, passive intent capture

Successfully integrating Supermemory for long-term cognition

Creating a clear, intuitive UI that demonstrates the idea effectively

What we learned

Memory is more powerful when it’s structured, not just stored

Passive tools can significantly reduce cognitive load

Selective capture is often more useful than recording everything

Persistent memory enables smarter, long-term experiences

What's next for NOVAMIND

Smarter categorization of sticky notes

Topic clustering across meetings and lectures

Calendar integration for time-bound notes

Multi-day and weekly summaries

Personal knowledge graphs built from remembered moments

Built With

• gemini-via-ai-sdk
• next.js-(app-router)
• react
• supermemory-(persistent-memory)
• tailwind-css
• typescript
• web-speech-api