一个专门用于从Effect生态系统获取最新文档的模型上下文协议（MCP）服务器。

目的

该 MCP 服务器为 AI 助手和工具提供实时访问 Effect 文档。它使 VS Code 扩展和其他 MCP 客户端能够检索准确且最新的 Effect 包文档，帮助开发者解答以下问题：

• 核心effect包
• 生态系统包（@effect/platform、@effect/ai等）
• 稳定和不稳定效果包

服务器智能地检测到代码中导入的 Effect 包，并获取最相关的文档来回答你的疑问。

特色

• Effect 生态系统的实时文档访问
• 支持稳定和实验性包
• 基于导入的上下文感知包检测
• 与VS Code及其他兼容MCP客户端的无缝集成
• 帮助解答Effect开发者的API使用问题

这个服务器暴露了什么

• 资源（优先）
• 动态模板：effect-docs://{libId}
• 支持{libId}的完成和资源列表

@niklaserik/effect-mcp

A Model Context Protocol (MCP) server dedicated to fetching the most current documentation from the Effect ecosystem.

Purpose

This MCP server provides real-time access to Effect documentation for AI assistants and tools. It enables VS Code extensions and other MCP clients to retrieve accurate and up-to-date Effect package documentation to assist developers with questions about:

• The core effect package
• Ecosystem packages (@effect/platform, @effect/ai, etc.)
• Both stable and unstable Effect packages

The server intelligently detects imported Effect packages in your code and fetches the most relevant documentation to answer your queries.

Features

• Real-time documentation access for Effect ecosystem
• Support for both stable and experimental packages
• Context-aware package detection based on imports
• Seamless integration with VS Code and other MCP-compatible clients
• Helps answer API usage questions for Effect developers

What this server exposes

• Resources (preferred)
  • Dynamic template: effect-docs://{libId}
  • Completions on {libId} and resource listing are supported
  • Static aliases for discovery: effect-docs://effect, effect-docs://@effect-platform, etc.
  • Supported libraries (as of now):
    • effect
    • @effect/platform
    • @effect/sql
    • @effect/vitest
    • @effect/ai
    • @effect/cli
    • @effect/cluster
    • @effect/rpc
    • @effect/typeclass
• Tools (for compatibility)
  • effect-documentation: fetch docs for selected libraries and return concatenated markdown
  • effect-doc-links: return plain URIs to the resources above so clients can follow them

Caching & Retry

• In-memory TTL cache (~10 minutes) to avoid refetching unchanged docs.
• Lightweight retry with small backoff on transient fetch errors.

Installation

pnpm add @niklaserik/effect-mcp

Usage

For usage in VS Code or other MCP clients, refer to the client’s documentation for adding custom MCP servers.

Quick test locally:

pnpm install
pnpm build
node dist/index.js # stdio transport

Then connect with an MCP client (e.g., MCP Inspector) over stdio and try:

• List resources and read effect-docs://effect
• Call tool effect-doc-links with { "libraries": ["effect", "@effect/platform"] }

Initialization instructions

• The server publishes short usage instructions via MCP initialize for clients that display them.

Development

# Install dependencies
pnpm install

# Build the project
pnpm build

# Run in development mode with auto-reload
pnpm dev

Notes:

• Built with @modelcontextprotocol/sdk@^1.17.4 using the modern registerResource/registerTool APIs.
• Transport: stdio by default; can be adapted to Streamable HTTP if needed.

License

MIT