Check_MateAI

Inspiration

We noticed a problem: searching for symptoms online often causes more anxiety than answers. “Dr. Google” isn’t always reliable. Over 16% of online results exaggerate risks, leading to “cyberchondria”, heightening stress and fear due to inaccurate online health information. Online symptom checkers often ignore personal context because they don’t know your medical history, prescriptions, or even that you hurt your back as a child. Yet, over 72% of U.S. adults look up health information online every year. This inspired us to create Check_MateAI, an AI-powered website that provides users with clear, concise, and personalized health insights.

What it does

Check_MateAI allows users to input symptoms and receive AI-driven responses in a clear, three-point format: urgency, suggestions, and next steps. But we didn’t stop there. To make the tool more accurate and personal, users can create an account to upload their medical records and prescriptions. This allows the AI to tailor its responses based on individual history, something “Dr. Google” could never do. We implemented voice input and output in order to make the website accessible to everyone and help users who struggle with typing, have vision issues, or are non-native speakers.

How we built it

We developed Check_MateAI using Next.js, TypeScript, and Tailwind CSS to create a fully responsive application. Users can provide symptoms via text or voice, in any language, and the app responds in the same language in real time. This was achieved by integrating ElevenLabs for speech-to-text and text-to-text, allowing seamless voice interaction, and the Gemini API for AI-powered symptom analysis based on users' history. Gemini processes the input to provide safe, non-diagnostic advice, including urgency, suggestions, and next steps. All sessions are stored in MongoDB, enabling users to view their history and receive personalized feedback.

Challenges we ran into

During development, we faced several challenges due to new technologies and the limited team size. Integrating ElevenLabs for real-time speech input and output in multiple languages was a first for us, requiring careful handling of voice recognition and text-to-speech synchronization with Gemini AI responses. Working with MongoDB for both of us was also new, and we had to learn dynamic data management to efficiently store, update, and retrieve user sessions and personalized history. Additionally, building a fully functional dashboard, AI integration, and voice features with a team of only two people meant managing a high workload and tight time constraints, especially during the hackathon. Despite these challenges, we successfully delivered a working MVP that combined text and voice input, multilingual support, and personalized health insights.

Accomplishments that we're proud of

We are proud of turning an idea into a fully functioning website that helps people experience less anxiety when searching for symptoms online. Users can now focus on what truly matters: understanding the urgency of their situation and taking actionable next steps. And we are more proud to finish our project on time and learn to use different new technologies in just 24 hours. Integrating Gemini and ElevenLabs transcriptions of text and speech approach in real time.

What we learned

Through building Check_MateAI, we learned the importance of accessibility, clarity, and responsible AI in healthcare. Creating a platform that accommodates diverse users, from the elderly to those with physical or language barriers, taught us that thoughtful design and inclusivity are as important as technology itself. Collaborating across different backgrounds, from computer science to architectural engineering, strengthened our teamwork and adaptability.

What's next for Check_MateAI

Our goal is to make Check_MateAI even smarter. In the future, we aim to collect anonymized user data and, with medical verification, use multifactor authentication to safeguard confidential medical data. Ultimately, we hope to help users make informed health decisions with confidence, reducing cyberchondria and improving mental well-being. We are proud to have built something that can help people find peace of mind instead of panic. Check_MateAI encourages users to focus on solutions and next steps rather than dwelling on worst-case scenarios.

Built With

• elevenlabs
• geminiapi
• mongodb
• next.js
• node.js
• tailwindcss
• typescript