Anchor

Home Page of the Device
User history of previous attacks for analysis and preservation
Example of Anchor helping the user by using a "grounding" technique
Example of Anchor helping the user by using a controlled breathing technique

Inspiration

Our team has a history of panic and anxiety attacks. While stigmas around mental health have improved, the gap between people in need and available resources still exists. During a panic attack, thinking about anything but dying is pretty hard. We wanted an app that could help relieve this by doing some of the heavy lifting and critical thinking for the user.

What it does

Anchor is an AI-powered companion that helps guide users through panic attacks in real time. It provides grounding exercises, visualized breathing techniques, and personalized video messages from loved ones. By integrating computer vision and conversational AI, Anchor adapts to the situation and delivers the right support when it’s needed most.

How we built it

We created a FastAPI backend that connects computer vision, user data, and Gemini API prompting with a Swift/SwiftUI front-end for iOS development. The system handles real-time interactions between vision-based insights and responsive AI-driven guidance, providing a seamless user experience.

Challenges we ran into

Running computer vision locally on the iphone/camera
Latency in communication between front-end and backend services.
Unfamiliarity with Apple’s ARKit made development slower.
Ensuring text-to-speech (TTS) and speech-to-text (STT) sounded natural and conversational.

Accomplishments that we're proud of

Building a sleek, realistic iOS integration and UI.
Tackling a problem deeply relevant to our team’s own experiences with mental health.
Successfully leveraging computer vision to strengthen AI prompting.

What we learned

We learned how to design for accessibility and emotional sensitivity in high-stress situations. We also gained hands-on experience combining iOS AR development, backend APIs, and computer vision pipelines into a real-time system. Beyond the technical challenges, we learned how important it is to align design decisions with empathy and user trust.

What's next for Anchor

Our next step is implementation with smart glasses, enabling hands-free, immersive support that feels natural in the moment. With wearable integration, Anchor can provide even more seamless, discreet, and personalized assistance.