Random Walk: A Tiny AI Nudge to Get Outside

Random Walk is a compact web application designed to encourage people to leave their homes and explore their surroundings. Users can select a brief walking duration of 5, 10, or 20 minutes, share their location with the browser, and the app generates a nearby waypoint, creates a walking directions link via Google Maps, and prompts artificial intelligence for a concise outdoor observation challenge.

This low-effort motivation tool is aimed at anyone seeking a simple push to step outside and enjoy their neighborhood. The minimalist interface relies on HTML, CSS, and browser JavaScript, while a small Node.js server handles the backend operations. The browser's geolocation feature is used solely to determine a random waypoint for the suggested walk; the coordinates are not transmitted to the AI or stored anywhere.

By integrating the GitHub Copilot SDK, the app can retrieve a short, safe observation activity tailored to a quick walk. This approach allows the mission generation to be performed on the server, while keeping user location data only within the browser, ensuring privacy and security. The open-source SDK enables the seamless integration of an agent runtime directly into the Node.js application, facilitating the mission generation process without relying on the Copilot Chat interface.

Participants in the Hacktoberfest Open-Source AI Challenge Week 1 were invited to submit innovative projects like Random Walk, competing for the Best Use of GitHub Copilot category.