peterlmajors/lox-genie
If you are the rightful owner of lox-genie and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
The Lox MCP Server is a Model Context Protocol server designed to integrate various tools and resources for fantasy football research and analysis.
Welcome to the Lox Genie
This repository provides access to the Lox Genie, a fantasy football research assistant built with LangGraph and open source language models (gpt-oss:20b, llama3.1:8b, etc.), as well as the Lox MCP Server, to offer fantasy football tools and resources in a unified format.
🧠 Thesis
The use case for an agentic fantasy sports consultant is compelling: access to sports data is heavily democratized, player performances and news are constantly updating, and 'expert opinion' is bountiful thanks to social media, but difficult to apply.
Lox is designed for the fantasy sports power-user. By building a context-rich relationship with each user, Lox is able to apply strategic preferences across leagues and seasons — remembering not just who you manage, but how you like to play.
Existing fantasy football sites like KeepTradeCut crowdsource valuations through direct player comparisons, while others like FantasyCalc apply optimization algorithms to real trades, generating market-driven rankings.
Behind the scenes, Lox performs a similar function, but with text data. Conversation histories are anonymized and useful insights are added to Lox's knowledge base, ensuring managers benefit from the network's collective intelligence. (Premium Feature)
Fantasy football brings people together. However, keeping up with the competition can turn into a time-consuming, isolating task. Let's fix that. Join Lox today.
🏈 What is Lox Genie?
Lox Genie is, fundamentally, a deep research agent which processes user queries by:
- Assessing their relevance for Lox's intended purpose and tool-calling cababilities
- Optionally, engaging a human-in-the-loop to gain additional information to achieve the user's goal
- Plans research strategies by breaking down complex questions into actionable subtasks
- Executes research using specialized MCP tools and resources for data gathering
- Provides expert analysis with concise, well-supported recommendations
The agent is instructed to be maximally truth-seeking, providing resolute and non-ambiguous answers by blending its knowledge base with ground-up analysis.
🏗️ Architecture
- FastAPI Service (Port 8000): RESTful API with streaming chat capabilities
- MCP Server (Port 8001): Model Context Protocol server for tool integration
- React Frontend (Port 80): Serves chat interface and curated news feed
- LangGraph Agent: Multi-node workflow orchestration for intelligent research
- Docker Support: Containerized deployment with docker-compose
🛠️ Prerequisites
- Python 3.11+
- uv (recommended) or pip
- Docker (for containerized deployment)
🚀 Quick Start
Using Docker (Recommended)
# Start both services
docker-compose up --build