Landing Page
Switch between focus modes
Dashboard showing metrics including Pattern recognition and Recall Information etc
Add/Update/Delete/Review Memories
Semantic Search for searching memories
Ask Recall about your memories
Ask recall to quiz about your memories

Inspiration

We were frustrated with passive memory tools that just store information. We wanted an AI assistant that actively helps you - detecting patterns, predicting what you'll forget, and telling you exactly what to do next based on your stored memories.

What it does

Rewind is a proactive memory intelligence platform that transforms your saved memories into actionable insights. It tracks deadlines, generates AI quizzes from your notes, detects learning patterns (best study times, knowledge gaps, retention trends), auto-saves conversations, and provides smart alerts about what needs your attention. It analyzes your habits and helps you stay on top of everything that matters.

How we built it

Built with Next.js 14 and TypeScript, leveraging Supermemory API for storage, Groq/Gemini for AI features, and SendGrid for digests. The frontend uses React 19, Tailwind CSS, and Recharts for data visualization. We designed specialized focus modes (Study, Job Search, Meeting) with contextual memory filtering and smart templates that auto-extract metadata from conversations.

Challenges we ran into

Designing an AI that's proactive without being annoying was tricky. We had to fine-tune alert algorithms to balance urgency detection with noise reduction. Implementing spaced repetition intelligence and pattern detection required complex memory analysis. Making the quiz generation contextually relevant and handling async Next.js 15 route parameters also posed challenges.

Accomplishments that we're proud of

Created 6 smart memory templates with automatic metadata extraction, built a pattern detection engine that discovers behavioral insights, implemented a retention scoring system that adapts to performance, and designed a timeline view showing when you learned something and when to review it. The chat auto-save feature intelligently detects learning content.

What we learned

Proactive AI requires careful UX design - timing and relevance matter more than feature count. We learned advanced Next.js patterns, memory management optimization, and how to build LLM pipelines that balance speed with quality. Pattern detection in user behavior is more valuable than simple reminders.

What's next for Rewind

Mobile app for on-the-go memory capture, voice note transcription with auto-categorization, collaborative memory spaces for teams, calendar integration for automatic meeting memory creation, and browser extension to capture web content directly. We also plan to add visual memory types (images, diagrams) and personalized spaced repetition algorithms.