Agent-FPL

Agent-FPL is a multi-agent AI framework that helps Fantasy Premier League (FPL) managers make smarter, data-driven decisions using live football data and Large Language Models (LLMs). The platform combines player statistics, fixture difficulty, injury updates, team news, and historical performance to generate intelligent, explainable transfer recommendations and squad insights. Unlike conventional chatbots, Agent-FPL adopts a collaborative multi-agent architecture where specialized AI agents handle distinct tasks such as retrieving live FPL data, analyzing player form, evaluating upcoming fixtures, validating player availability, assessing budget constraints, and recommending optimal transfers. This division of responsibilities enables more accurate, transparent, and context-aware decision-making. To improve reliability, the framework implements Retrieval-Augmented Generation (RAG). Football knowledge, player information, and statistical datasets are indexed in a vector database, allowing semantic search to retrieve relevant context before the LLM generates a response. This approach grounds recommendations in current data, reducing hallucinations and improving factual accuracy. Every recommendation includes a clear explanation based on recent form, fixture analysis, player statistics, ownership trends, and potential risks, helping users understand why a transfer is suggested rather than simply presenting a ranked list. The modular architecture allows new capabilities such as captain recommendations, wildcard planning, chip strategy optimization, and personalized squad analysis to be added with minimal changes. Agent-FPL showcases modern AI engineering concepts including multi-agent orchestration, prompt engineering, RAG, vector search, semantic retrieval, API integration, and structured data processing. Although developed for Fantasy Premier League