๐ About the Project: Vibe Learning
"Cursor for Learning" โ Personalized, AI-powered education through mindmaps, roadmaps, and practical projects.
๐ Inspiration
During my own learning journey in AI and development, I constantly found myself switching between YouTube videos, online notes, ChatGPT prompts, and personal notes. Traditional platforms felt either too rigid or too fragmented. I wished for a tool that could:
- Structure my learning like a mindmap
- Adapt content to my pace and skill level
- Let me interact with knowledge using natural language
- Help me build projects along the way
That's how Vibe Learning was born โ a learning platform that bridges the gap between AI tutoring, personalized content, and hands-on project building.
๐ ๏ธ How I Built It
Vibe Learning is powered by a full-stack AI-integrated system:
๐งฉ Tech Stack
- Frontend:
Next.js 15,Tailwind CSS,Radix UI - Backend:
tRPC,Prisma ORM,PostgreSQL,Supabase - AI Engine:
Google Gemini 2.0 Flash, fallback to 1.5 Flash - PDF/Document AI: Python backend using NLP + LangChain-like tools
- Visualization: Custom
Mindmap.tsxwith zoom, pan, and color-coded nodes - Video Curation: YouTube Data API v3 for top-ranked videos per topic
- Command Interface: A conversational parser with
/explain,/visualize,/compare, and more
๐ Core Features
- AI-Powered Learning Roadmaps: Auto-generated from a single topic
- Interactive Mindmaps: Visualize and navigate learning hierarchies
- Conversational Commands:
/explain @topic,/visualize,/compare - Project Generator: Builds hands-on projects aligned to learning goals
- Progress Tracking: User libraries and resumption from last session
๐ What I Learned
- How to design AI-first interfaces that go beyond chat
- Building scalable architectures with tRPC and Prisma
- Handling natural language command parsing
- Integrating multi-modal learning: visual + conversational + practical
- Managing AI model limitations with fallback strategies and user feedback loops
๐ง Challenges Faced
- Contextual AI Parsing: Designing a command parser that understands fuzzy and chained inputs
- Mindmap Scalability: Managing dynamic node sizes, levels, and zoom performance
- AI Resource Quality: Filtering low-quality videos and ensuring relevance
- Multi-Difficulty Generation: Balancing content complexity across beginner, intermediate, and advanced levels
- PDF Understanding: Extracting and summarizing large document content reliably with limited context window
๐ฎ Whatโs Next
- ๐ง Add
/solveand/analyzefor PDF problem-solving and pattern recognition - ๐ฌ Add user-specific memory to track long-term learning progress
- ๐ Launch a curated library of beginner-friendly AI courses
- ๐ Integrate spaced repetition and quiz generation features
- ๐ฑ Build a mobile-first interface for micro-learning
โค๏ธ Final Thoughts
Vibe Learning is more than a project โ itโs the platform I wish I had while learning. If it helps even one learner find clarity, structure, or inspiration, thatโs a win.
Feel free to explore, contribute, or just vibe with learning.
Built With
- 15
- 2.0
- ai
- api
- cloud
- command
- css
- custom
- data
- eslint
- flash)
- gemini
- langchain-style
- lucide
- mermaid.js
- natural-language-processing
- next.js
- orm
- parser
- pipeline
- postgresql
- prettier
- prisma
- python
- radix
- react
- roadmap-structuring
- sdk
- sql
- supabase
- tailwind
- trpc
- typescript
- ui
- v3
- vercel
- youtube


Log in or sign up for Devpost to join the conversation.