Doubt Solver

Real-Time Learning with WebRTC, Socket.IO & AI

The Solution

Built a real-time platform where students upload image-based doubts. The system decides between AI-generated solutions (using Gemini API) or connecting the student to a live tutor for immediate assistance.

Core Features

  • AI Image Understanding: Integrated Gemini API to analyze uploaded images and provide contextual text-based solutions.
  • Live Tutor Matching: Real-time session coordination and tutor matching using Socket.IO.
  • 1:1 Video Calls: High-bandwidth, low-latency video communication enabled by WebRTC.
  • Collaborative Whiteboard: Custom HTML Canvas implementation for visual problem solving during live sessions.
  • Mutual Rating System: Post-session feedback loop for quality assurance between students and tutors.

Technical Challenge

Low-Latency Communcation: Syncing a whiteboard and video call simultaneously across different network conditions.

Solution: Leveraged WebRTC for the data channel to sync canvas coordinates in real-time while maintaining stable video streams through STUN/TURN servers.

The 60-Second Pitch

"I developed Doubt Solver to modernize student-tutor interactions. It’s a real-time platform where I combined WebRTC for video calls and Gemini API for automated image-based doubt solving. I also built a collaborative whiteboard using HTML Canvas to make learning interactive. This project gave me deep experience in handling complex real-time events and multi-user synchronization."