A Large-Scale-Dataset-of-MCP-Implementation-on-GitHub

The Model Context Protocol (MCP) Dataset provides a large-scale, evidence-based analysis of open-source MCP implementations on GitHub.This dataset captures how developers integrate, extend, and maintain the MCP standard across clients, servers, and gateway architectures. The collection process involved: Automated discovery of 2,297 candidate repositories via the GitHub REST Multi-stage validation to remove false positives and confirm MCP evidence through manifest files, configuration markers, and commit-level traces Manual review for precision estimation (83% ± 4% at 95% confidence) Comprehensive metadata including repository-level activity (commits, pull requests, contributors, stars), dominant programming language, and classification into client, server, gateway, or hybrid designs Each verified repository is represented in structured JSONL supplemented with scripts, schema documentation, and figures.The dataset follows FAIR principles: Findable, Accessible, Interoperable, and Reusable. This serves as a foundation for reproducible research on emerging AI infrastructure standards.