Inspiration Every student knows the frustration of asking an AI chatbot for study help, only to get robotic textbook definitions, instant homework spoilers, or overwhelming walls of text. We built MindMate to turn AI from a passive answer generator into a genuine personal study coach—one that actually understands where you are stuck, sticks strictly to your course syllabus, and guides you step-by-step so you truly learn.
What it does MindMate is an intelligent study companion that reads lecture slides, textbook PDFs, and diagrams, organizing them into an interactive visual roadmap of connected topics. Instead of just giving away answers, it coaches you through friendly guided conversations, offers progressive hints, and tests your understanding with interactive "Knowledge Detective" challenges. It tracks your learning progress on every single topic in real time and automatically creates a custom daily study plan focused on your weak spots.
How we built it We built MindMate with a modern, responsive web interface using React and Tailwind CSS, powered by a fast Python backend. We integrated Google Gemini to analyze lecture slides and visual diagrams with high accuracy. We used interactive visual graphing tools to map out how course topics link together, and added live progress charts so students can clearly watch their confidence and understanding grow with each study session.
Challenges we ran into Our biggest challenge was stopping the AI from making up facts or immediately blurting out the final answer. Standard chatbots default to giving away complete solutions right away. We designed smart guardrails that lock the AI strictly to the student's uploaded class materials and built a structured hint ladder that nudges students in the right direction without spoiling the solution.
Accomplishments that we're proud of We are proud of creating a reliable study companion that has passed 110 automated tests and consistently stays true to syllabus material without hallucinating. Seeing the visual concept roadmap, adaptive daily planner, and conversational tutor work together to help students bridge their learning gaps was our most rewarding milestone.
What we learned We learned that great learning technology isn't about providing instant answers—it is about asking the right questions at the right time. Encouraging students to explain ideas in their own words and guiding them through hands-on practice builds far deeper comprehension and long-term memory than reading passive text summaries ever could.
What's next for MindMate Next for MindMate is a real-time voice conversation mode so students can talk through complex problems out loud naturally. We also plan to add collaborative study rooms where classmates can team up to solve practice cases, along with one-click integrations for school portals like Canvas and Blackboard to automatically sync assignments and lecture notes.
Built With
- bayesian-knowledge-tracing
- docker
- fastapi
- google-gemini
- groq
- jwt
- langchain
- lucide
- numpy
- openai
- pgvector
- postgresql
- pymupdf
- python
- rag
- react
- render
- rest-api
- supabase
- tailwind-css
- typescript
- vercel
- vite
- whisper
Log in or sign up for Devpost to join the conversation.