Inspiration
Being in tech, it’s so easy to lose track of health and fitness — long hours, irregular meals, and no time for the gym. I wanted to fix that. The idea was to build something simple, free, and efficient — an AI fitness coach that helps people like us stay consistent without overcomplicating things.
What it does
It’s an AI-powered health and fitness coach that generates personalized workout and meal plans in seconds. You can chat with it, tell it your goals — like “I want to gain muscle” or “I need a high-protein vegetarian plan” — and it instantly builds daily routines using AWS Bedrock and Spoonacular API.
How we built it
We built the backend using AWS Lambda functions integrated with Bedrock for AI generation. The frontend runs on a lightweight local server that talks directly with these Lambda functions. Everything is connected, the workout generator, meal planner, and fitness chat assistant — all running serverlessly and efficiently.
Challenges we ran into
Getting Bedrock to return clean, valid JSON (AI models love to talk too much). Managing AWS Lambda timeouts during model calls. Keeping the setup simple enough for anyone to deploy and use for free.
Accomplishments that we're proud of
Fully functional AI coach that actually builds personalized routines. Smooth integration of multiple Lambda functions and Bedrock. Keeping it all serverless and free to run.
What we learned
AI models can be unpredictable: structure and validation are everything. Also, simplicity wins. You don’t need a huge system to build something genuinely useful.
What's next for AI Health and Fitness Coach
Add progress tracking and smart recommendations over time. Build a small app version for mobile users. Open it up so others can use it to build their own AI fitness setups.
Built With
- amazon-web-services
- pydantic
- python
- render
- spoonacularapi

Log in or sign up for Devpost to join the conversation.