Euryalus

Logo slide

Euryalus - Your Personal AI Endurance Coach The Inspiration: Making Elite Coaching Accessible As avid endurance athletes, we've always been fascinated by the science of training, recovery, and peak performance. However, we quickly realized that access to high-quality, personalized coaching is a luxury. A personal coach can cost upwards of $4,000 a year, putting it out of reach for most amateur athletes. At the same time, we saw millions of athletes on platforms like Strava—over 150 million, in fact—meticulously tracking their activities but lacking the guidance to interpret that data and train effectively.

This gap was our inspiration for Euryalus. We wanted to create a tool that could democratize endurance coaching, making personalized, intelligent, and affordable training advice accessible to everyone. The goal was to build a virtual coach that understands your workouts, your body, and your goals, helping you train smarter, recover better, and avoid injury.

How It Works: A Look Under the Hood

Euryalus is designed to be a seamless addition to an athlete's existing routine. The technical architecture is built around a few key components that work together to deliver personalized coaching.

Here is the data flow:

User Interaction: The athlete interacts with a simple front end built with HTML, CSS, and JavaScript. Here, they can view their workout calendar, provide feedback, and input information about any injuries or how they're feeling.

Backend Processing: This user data is sent to a Node.js backend server, which acts as the central brain of the operation.

Data Integration: The backend connects to the Strava API to pull the user's workout data—runs, rides, swims, and more. It also gathers personal info and any injury details provided by the user.

The AI Core: This is where the magic happens. All the processed data—workout stats, user feedback, injury info—is compiled into a carefully engineered prompt and sent to the Google Gemini AI.

Personalized Feedback: Gemini analyzes the comprehensive data and generates targeted workouts, recovery advice, and even words of encouragement, tailored specifically to that athlete's recent performance and current state. This feedback is then sent back through the backend and displayed to the user on the front end.

Challenges we ran into

Building Euryalus was a journey filled with learning opportunities, disguised as challenges:

Prompt Engineering: The biggest challenge was crafting the perfect prompt for the Gemini AI. The quality of the AI's output is directly proportional to the quality of the input. It took countless iterations to design a prompt that consistently delivered safe, effective, and motivating coaching advice, rather than generic or potentially harmful suggestions. We had to learn how to structure the data and frame the questions to the AI to get expert-level results.

API Integration: Juggling multiple APIs (Strava for data input, Google Cloud for hosting, and Gemini for intelligence) was complex. Each has its own authentication methods, rate limits, and data formats. Ensuring they all communicated seamlessly required careful planning and robust error handling.

Data Privacy and Trust: Since we're handling personal health and workout data, ensuring user privacy and building trust was paramount. This meant implementing secure data handling practices from day one and being transparent about how the data is used.

What we learned

This project was an incredible learning experience. We dove deep into full-stack development, from creating an intuitive user interface to building a scalable backend on Google Cloud. More importantly, we learned how to harness the immense power of Large Language Models like Gemini to create truly personalized and dynamic applications.

The journey for Euryalus is far from over. The vision is to expand its capabilities significantly. Some of the features on the roadmap include:

Expanding to More Sports: Applying the core technology to a wider range of athletic activities.

Google Calendar Integration: Allowing Euryalus to plan workouts around a user's busy work or school schedule.

Custom Playlists: Generating playlists that match the intensity and duration of planned workouts.

AI-Powered Video Analysis: The most ambitious next step is to incorporate video analysis to give athletes feedback on their form and technique, providing a new level of tailored improvements and rehabilitation.

Euryalus started as an idea to solve a problem we experienced firsthand, and it has evolved into a powerful platform with the potential to help millions of athletes achieve their goals.