Study-Flow

Your intelligent study companion. Track progress, collaborate with peers, and unlock your academic potential with AI-powered insights.

Comment

Inspiration

Built StudyMate during mid-semester exams to solve a problem I faced daily – managing study time effectively while juggling academics. The irony? Building a study tool while supposed to be studying. But it taught me more about time management than any productivity app ever could! MentorMind's first advice? "Take breaks more often" – advice I clearly wasn't following while coding at 2 AM.

What it does

Study-Flow is an AI-powered study management platform featuring:

  • Smart Study Timer: Customizable Pomodoro sessions with automatic tracking and Spotify integration for focus music
  • MentorMind AI Assistant: Multi-agent system (GPT-OSS 120B, Nova Lite, Claude 3.5 Haiku) with RAG architecture for personalized study advice
  • Auth0 Integration: Secure third-party API access via Token Vault (Google Calendar & Spotify live)
  • MCP Tool Support: Automatic external resource fetching – just drop a URL and MentorMind reads it with you
  • Performance Analytics: Visual dashboards tracking study patterns and progress
  • Study Groups: Collaborative workspaces with real-time chat and leaderboards
  • Coin Economy: Gamification system (earn 1 coin/second studying, spend 100 coins/AI query)

How we built it

Tech Stack:

  • Next.js 14 + TypeScript frontend with Convex real-time database
  • Hybrid auth architecture: Convex Auth for users + Auth0 Token Vault for API tokens
  • Heroku Managed Inference powering multi-agent AI system
  • MCP integration with dynamic tool discovery and SSE parsing

Key Innovation: Hybrid auth pattern (undocumented elsewhere) – kept Convex Auth for familiar user login, added Auth0 exclusively for third-party API token management. Zero downtime migration with feature flags.

Code Impact: Auth0 Token Vault reduced token management from 120 lines of manual refresh logic to 30 lines of enterprise-grade automation. That's 75% code reduction while upgrading to SOC 2 Type II certified security.

Challenges we ran into

  • MCP Integration Complexity: SSE parsing, dynamic tool routing, and understanding Heroku's Agents API took days to crack. The breakthrough came from building a custom SSE parser and provisioning all available tools to the AI.
  • OAuth Understanding: Confused ID tokens, access tokens, and refresh tokens initially. Auth0's docs clarified the differences beautifully.
  • Hybrid Auth Architecture: No documented pattern for combining Convex Auth + Auth0 Token Vault. Designed it from scratch with automatic fallbacks.
  • Balancing Development: Building during exams meant 3-hour sleep nights. Used MentorMind to help me study while building it (meta!).

Accomplishments that we're proud of

  • Production-Ready Infrastructure: Heroku-deployed MCP server with 10 working calendar tools, multi-user isolation, rate limiting
  • Hybrid Auth Innovation: Pattern allowing zero-downtime migration between auth systems with feature flag control
  • Spotify Integration Live: Full playback control in study timer with Auth0-secured token management, configurable via settings
  • Collaborative AI Deployed: Shared MentorMind assistants for study groups with full group context and member insights
  • Meta Achievement: Used MentorMind to study for exams while building MentorMind. Got B+ on exam I was supposed to fail!
  • Real Impact: Context-aware AI that checks actual Google Calendar to suggest optimal study times

What we learned

  • Don't Reinvent OAuth: Auth0 Token Vault solved token management better than any custom implementation. Just use proven solutions.
  • Hybrid > All-or-Nothing: Keep what works, add what you need. No need to rip out existing auth.
  • Feature Flags Are Magic: Testing Auth0 alongside Convex system gave confidence for safe migration.
  • Good Documentation Matters: Auth0's AI agents docs and Heroku's MCP GitHub examples were excellent.
  • Real Data Testing Reveals Truth: Mock tests are fine, but testing with actual Google Calendar revealed edge cases we'd never find otherwise.

What's next for Study-Flow

  • Email/Gmail Integration: Track email-based assignments and deadlines
  • Notion Sync: Import study notes for better AI context
  • Voice Interaction: Hands-free studying for memorization sessions
  • Mobile App: On-the-go study tracking and session management
  • Advanced Spotify Features: Automatic playlist switching based on study intensity and break detection
  • AI-Powered Study Plans: Auto-generated weekly schedules optimized for exam prep

Built with: Claude Code (frontend logic), Convex Chef (backend development), TweakCN (theme customization), and Aceternity UI (meteor animations & hover effects)

Built With

Share this project:

Updates