nanbingxyz/mcpsvr

What it solves

MCPSvr provides a centralized, community-driven directory for discovering and reviewing Model Context Protocol (MCP) servers. It addresses the fragmentation of MCP server discovery by providing a shared registry that allows both humans to evaluate tools and MCP clients to automate the installation and setup of these servers.

How it works

The project is a lightweight web directory built with Next.js. It centers around a curated servers.json registry file that contains metadata for various MCP servers, including execution commands, runtime arguments, environment variables, and documentation links. It uses a specific schema for server entries and a placeholder system for user-provided inputs (e.g., {{paramName@paramType::paramDescription}}) to facilitate direct installation by compatible MCP clients.

Who it’s for

Developers who need to find and compatible MCP servers to extend the AI agent capabilities of their LLM clients, as well as contributors who wish to maintain a curated list of of MCP servers.

Highlights

• Centralized Registry: A single, browsableable directory of MCP servers.
• Automated Setup: Metadata provided in the registry allows compatible MCP clients to install and run servers directly.
• Community-Driven: New servers and more accurate metadata can be added via pull requests.
• Standardized Schema: Uses a structured JSON schema to ensure consistency across entries.