LegendAIry (Rift Rewind Hackathon)

Be a legend, make it daily, powered by AI. LegendAIry breaks down LoL matches, showing why you die, your patterns, and how to improve

Comment

Inspiration

MEXICO IS EXLUDED, so I can’t continue my submission, which is unfortunate so i will not send my job done.

As both a gamer and music creator, I’ve always been captivated by how AI can transform the gaming experience. One of the most frustrating moments in League of Legends is dying repeatedly without truly understanding why. That pain sparked the idea for Legend.AIry:

What if AI could analyze your matches, explain the reasons behind each death, and teach you how to improve? My background in music also shaped the vision: I wanted the analysis to feel, not just inform — blending data, visuals, and sound into an immersive and emotional experience.

What it does

Legend.AIry uses match data from the Riot Games API, combined with AI reasoning via Amazon Bedrock, to deliver deep, human-like insights:

Explains why you die in each match

Highlights recurring patterns like overextending, lack of vision, or chain CC

Provides actionable advice for improvement

Visualizes death events on an interactive timeline

Enhances the experience with original music and dynamic visuals that react to your data

How i built it

Data ingestion: Integrated the Riot Games API to fetch match_id, death events, and contextual information. AI reasoning: Created Python logic to classify deaths, then used Bedrock LLMs for natural language explanations.

def classify_death(event): if event['vision'] == 0: return "No vision" elif event['cc_chain'] > 1: return "Chain CC" else: return "Overextend"

AWS integration: Automated pipelines with Lambda and SNS for real-time alerts and scalability. UX/UI design: Built mockups for dashboards like “Why You Die,” the interactive timeline, and a weekly improvement panel. Multimedia demo: Synchronized original music and visuals to transform gameplay analytics into an immersive narrative experience. Challenges I ran into

Managing Riot API rate limits while collecting meaningful data

Translating raw logs into insights players can instantly understand

Refining prompt engineering for concise and accurate AI feedback

Building a cohesive UX using free prototyping tools

Handling all aspects solo coding, design, narrative, and music

Accomplishments that we're proud of

Successfully connected Riot Games API with AWS Bedrock for intelligent match analysis

Built a solo prototype merging gameplay data, AI reasoning, and music

Generated clear, actionable insights based on real player behavior

Designed engaging UI mockups and a multimedia demo ready for hackathon presentation

What I’ve learned

Applying LLMs for contextual reasoning in gameplay data

Designing user-centered interfaces for data-heavy environments

Integrating AWS cloud services (Lambda, SNS, DynamoDB) into a seamless pipeline

The importance of iterative prototyping starting simple and scaling smart

What's next for LegendAIry (Rift Rewind Hackathon)

Expand the AI model to support team-based analysis and synergy metrics

Develop real-time feedback during live matches

Refine the UX/UI for mobile and cross-platform use

Integrate interactive music layers that evolve with gameplay data

Prepare a complete, immersive demo combining AI, data visualization, and original sound design

Built With

Share this project:

Updates