MetaDraft AI

AI-powered SaaS to help League of legends teams optimize drafts: manage champ pools, analyze picks/bans, and get smart suggestions for scrims or matches. Dominate draft, win games.

Comment

Inspiration

As passionate players and analysts of League of Legends, we saw how crucial the draft phase is in determining the outcome of high-level games. While pro teams spend hours preparing for pick/ban phases, there was no centralized tool to track data, manage champion pools, and support decision-making in a structured, AI-assisted way. This inspired us to create MetaDraft AI — a platform built to help teams turn drafting into a real competitive edge.

What it does

MetaDraft AI is a SaaS platform designed for competitive League of Legends teams. It allows coaches and analysts to:

  • Manage player champion pools and roles with mastery levels
  • Simulate drafts with real-time pick/ban interactions
  • Import/export draft data from csv professional db
  • Analyze win/loss history, matchups, and trends
  • Receive AI-powered suggestions based on team style, patch data, and opponent tendencies

How we built it

We built MetaDraft AI using:

  • Frontend: React with Tailwind CSS and Vite for a modern, performant UI
  • Backend: Supabase for authentication, storage, and Edge Functions
  • AI integration: OpenAI API (GPT-4o) for draft suggestions and matchup analysis
  • Data ingestion: Custom parsers for JSON/CSV and Leaguepedia HTML data

Challenges we ran into

  • Parsing and standardizing draft data from different sources like Leaguepedia was complex due to inconsistent formatting, and finally found a well parsed csv file that contain all data I need...
  • Designing a UI that feels intuitive to coaches while handling complex interactions like simultaneous picks and bans.
  • Balancing AI suggestions without overfitting to public winrates or unreliable matchup data.

Accomplishments that we're proud of

  • Built a functional prototype capable of simulating and recording drafts
  • Developed a clean data model for teams, players, champion pools, and matches
  • Integrated AI-driven recommendations that adapt to patch and meta
  • Made it easy to onboard new teams with import/export and player pool management

What we learned

  • How to build a real-world SaaS around esports workflows
  • The complexity behind pro-level drafting and how every team has unique preferences
  • One of the biggest challenges during the draft phase is delivering the right information at the right time. We learned that providing too much data too fast can overwhelm decision-makers, especially under time pressure. The key is to find the right balance between speed, accuracy, and relevance, since drafts are timed and every second matters.
  • How to structure flexible data systems that support fast queries, custom filters, and AI feedback

What's next for MetaDraft AI

  • Add team-vs-team draft simulations based on opponent history
  • Build a scrim draft mode with dual-team collaboration
  • Develop privacy features for confidential drafts and scrims
  • Introduce a premium tier with live AI recommendations and scouting tools
  • Partner with LFL/ERL teams for beta testing and feedback
  • Define a user segment for regular players looking to get better at drafting.

Built With

Share this project:

Updates