Inspiration Tech support is currently "written in alien" for over 800 million people who struggle with basic digital tasks. We were inspired by family members who feel helpless when the "internet is broken" or pay €80 for a "two-click" fix. We wanted to build an assistant that feels like a patient friend, not a manual.

What it does EasySupport helps non-tech-savvy users diagnose everyday problems like slow laptops, printer errors, or "black screens". It removes all jargon—replacing "driver" with "helper program"—and guides users one step at a time. Users can type, talk, or point their camera at a device to get help.

How we built it Frontend: React with a focus on accessibility (high contrast, large text).Backend: FastAPI for high-performance AI orchestration.AI: Gemini API for jargon-free reasoning and vision-based diagnosis.Data & Cloud: MongoDB for session context and Digital Ocean for scalable hosting.Web3: Solana to issue "Proof of Fix" tokens to reward user learning.

Challenges we ran into The hardest part was "jargon-proofing" the AI to ensure it never says things like "DNS" or "IP address". We also had to ensure the "Show Me" mode could accurately identify physical ports and screen errors via camera in real-time.

Accomplishments that we're proud of We built a system that prioritizes privacy by design—no remote access and no personal data storage. Our "grandma-proof" logic confirms every step before moving forward, ensuring the user never feels overwhelmed.

What we learned We learned that technical accuracy matters less than empathy. Acknowledging a user's frustration and providing warm, encouraging feedback is what truly helps them solve the problem.

What's next for EasySupport AR Guidance: Overlaying glowing arrows on the camera feed to show exactly where to plug in cables.Auto-Diagnosis: Running background checks for disk space or internet speed and explaining them in plain language.Multi-language: Expanding support to all major European languages.

Built With

Share this project:

Updates