Agent Sportacus

System Architecture
Login/Sign Up
Onboarding/Profile
Dashboard
Journal Dashboard

Inspiration

We wanted to create a personalized fitness companion that could understand each user's unique goals, limitations, and preferences. Traditional fitness apps either provide generic workouts or simple tracking without intelligent guidance. We saw an opportunity to use AI to bridge this gap - creating a system that could have natural conversations about fitness while providing truly personalized recommendations based on individual user profiles.

What it does

Agent Sportacus is an AI-powered fitness trainer that provides personalized workout plans and nutrition advice through natural conversation. Users complete a comprehensive fitness profile including their goals, experience level, available equipment, and health conditions. The AI then uses this information to generate customized workouts and provide tailored fitness guidance. Users can also track their workouts in a digital journal, with the AI able to automatically parse and save workout details from conversational descriptions.

How we built it

We built the application using FastAPI for the backend with Amazon Bedrock powering the AI conversations. The frontend uses vanilla HTML, CSS, and JavaScript for simplicity and performance. We created a SQLite database to store user profiles, workout entries, and exercise data. The AI agent receives user profile context with each request to ensure personalized responses. We implemented JWT authentication for secure user sessions and designed the interface to work seamlessly across desktop and mobile devices.

Challenges we ran into

Our biggest challenge was learning AWS and Amazon Bedrock from scratch, as we had no prior experience with these services. We spent considerable time understanding how to configure Bedrock agents, manage credentials, and integrate AI responses effectively. We also struggled with creating reliable natural language parsing to convert workout descriptions into structured data, and ensuring the AI provided safe, appropriate fitness advice based on user limitations and health conditions.

Accomplishments that we're proud of

We successfully created a fully functional AI fitness trainer that provides genuinely personalized advice based on comprehensive user profiles. The seamless integration between conversational AI and workout tracking allows users to naturally describe their workouts and have them automatically saved. We built a clean, responsive interface that works well on both desktop and mobile devices, and implemented secure user authentication with proper data isolation between users.

What we learned

We gained valuable experience with AWS services, particularly Amazon Bedrock and how to effectively prompt AI models for personalized responses. We learned the importance of context-aware AI interactions and discovered that users prefer natural conversation over rigid command structures. We also learned key principles of building secure web applications with proper authentication and data validation, and the value of starting with core functionality before adding advanced features.

What's next for Agent Sportacus

We plan to add progress analytics with visual charts showing workout consistency and strength improvements over time. Integration with wearable devices and fitness tracking apps would provide more comprehensive health insights. We'd like to implement exercise video demonstrations and form guidance for safety. The introduction of a RAG system and additional features could include meal planning integration, social challenges with other users, and voice interaction for hands-free workout logging during gym sessions.