FortunAI — Miss Fortune Coach

FortunAI is a champion-specific AI coach for League players, starting with MF! Built with Riot API + AWS, it delivers tailored gameplay insights and paves the way for a full roster of AI mentors.

Comment

Inspiration

FortunAI began from my curiosity about whether League of Legends players could get champion-specific coaching instead of the generic statistics found on sites like op.gg. As a long-time Miss Fortune main, I saw that most trackers stop at numbers and never explain why those numbers matter. I wanted to create something that helps players reflect, learn, and celebrate their growth through context-aware insights that also capture the champion’s personality.

What it does

FortunAI is a champion-specific AI coaching assistant for League of Legends players, starting with Miss Fortune as the first supported champion. It retrieves match data through the Riot Games API and analyzes gameplay patterns that matter most to her — farming efficiency, early survivability, and presence around dragon objectives.

These metrics were intentionally selected to reflect Miss Fortune’s lane-dominant playstyle and early power spikes, allowing the coach to evaluate how effectively a player is using her strengths. The goal is to turn gameplay data into meaningful feedback that helps players make smarter decisions and master Miss Fortune’s mechanics, rather than just tracking generic stats.

How I built it

I built the backend using FastAPI and deployed it on AWS Lambda with Amazon API Gateway for public access. The system retrieves a player’s 25 most recent Miss Fortune matches through the Riot Games API and processes them in Python. This capped sample keeps analysis lightweight and efficient while still providing enough data for consistent, champion-specific insights.

I designed the analysis layer specifically around Miss Fortune’s gameplay. The model tracks champion-relevant metrics such as CS at 10 minutes, deaths before mythic completion, and dragon fight presence, which reflect how well a player is using her early power spikes and objective control. These curated metrics let FortunAI provide advice that is genuinely relevant to Miss Fortune’s strengths, weaknesses, and tempo.

The results are formatted into tailored feedback that matches her tone and personality. The stack includes Amazon Polly for future voice output and is structured to connect with Amazon Bedrock for natural-language coaching insights in later iterations.

Challenges I ran into

Key challenges included handling Riot API rate limits and maintaining JSON-parsing stability within AWS Lambda’s memory limits. Designing advice that felt authentic to Miss Fortune while staying objective also required careful tuning. I spent time balancing analysis speed, data size, and cost to keep the experience lightweight and reliable.

Accomplishments that Im proud of

I built a live, publicly accessible API that delivers actionable coaching feedback tailored to one champion’s mechanics. I’m proud that the entire stack is serverless, fast, and inexpensive to run. FortunAI proves that AI-assisted coaching can be specific, personality-driven, and educational—something no generic stat site offers today.

What I learned

I learned how to connect game analytics with natural-language feedback to create a coaching experience that feels personal rather than statistical. Working with AWS taught me how to use managed services effectively for quick, scalable deployment. I also learned that focusing on a single champion first leads to more precise insights and a clearer user experience.

What's next for FortunAI — Miss Fortune Coach

The next step is integrating Amazon Bedrock to generate dynamic coaching dialogue that sounds like Miss Fortune giving direct feedback. After that, I plan to expand FortunAI into a roster of champion-specific mentors. The long-term vision is a network of interactive agents that help players reflect, learn, and celebrate their progress across every role.

Built With

Share this project:

Updates