FixIT

FixiT Inspiration

Electronic devices are an important part of our daily lives, but it is often difficult to know whether visible damage actually needs repair. Many people spend money on unnecessary service visits, while others ignore minor issues that later become expensive to fix. We wanted to build an AI-powered assistant that gives users an instant and honest visual inspection using only their smartphone camera, helping them make informed decisions before visiting a repair center.

What it does

FixiT is an AI-powered visual inspection app that analyzes images of damaged electronic devices. Users simply point their camera at a device or upload a photo, and the app identifies visible damage, highlights the affected area, explains the issue in simple language, and suggests the next steps. If no damage is detected, the app reports "No Visible Issue" instead of making uncertain claims, making the results more reliable and trustworthy.

How we built it

We built the mobile application using React Native and Expo, creating a clean and responsive interface with camera access, gallery support, haptic feedback, and smooth navigation.

The backend was developed using Node.js and Express, handling image uploads, image compression, and communication with Gemini Vision.

To improve reliability, we implemented structured JSON prompting and schema validation so every AI response follows a consistent format. We also developed a visual overlay system that accurately highlights damaged regions on different screen sizes and implemented reliable API communication with retries and error handling. Finally, we deployed the backend on Render and packaged the application as a working Android APK.

Challenges we ran into

Our biggest challenge was reducing false positives. Shadows, reflections, and poor lighting occasionally caused the AI to mistake normal surfaces for damage.

To solve this, we redesigned our prompts and validation rules so the AI now defaults to "No Visible Issue" unless there is strong evidence of physical damage, such as cracks, corrosion, dents, or burn marks.

Another challenge was ensuring smooth communication between the mobile app and backend while maintaining a responsive user experience, even under unstable network conditions.

Accomplishments that we're proud of

We successfully built a complete AI-powered mobile application from concept to deployment. We integrated Gemini Vision for structured visual diagnostics, developed a visual damage-highlighting system, significantly reduced false positives through prompt engineering and validation, and deployed the backend while testing the Android APK successfully on a real device.

What we learned

Building FixiT taught us that creating reliable AI applications requires much more than connecting an AI model to an interface. Prompt engineering, response validation, backend architecture, user experience, and safety rules are all equally important in delivering trustworthy AI. We also gained valuable experience in mobile app development, cloud deployment, computer vision integration, and designing AI-powered user experiences.

What's next for FixIT

Our next goal is to expand FixiT into a complete AI repair assistant. We plan to support more device categories, estimate repair costs, provide repairability scores, recommend nearby repair services, add multilingual support, and enable richer AI conversations that guide users through troubleshooting step by step.

Ultimately, we want FixiT to become the first place people turn to when they want to understand a device issue before seeking professional repair.