graphql2mcp

graphql2mcp

Convert GraphQL schemas and endpoints into Model Context Protocol (MCP) servers.

Category
访问服务器

README

graphql2mcp

npm version npm downloads TypeScript license docs

Convert GraphQL schemas and endpoints into Model Context Protocol (MCP) servers. Point at any GraphQL API and get an MCP server with tools mapped from queries and mutations.

Features

  • Library mode — add GraphQL-backed tools to your existing TypeScript MCP server with one function call
  • Zero config proxy — pass a GraphQL endpoint URL and get an MCP server with every query as a tool
  • Mutation control — expose all mutations, none, or an explicit whitelist
  • MCP tool annotations — queries get readOnlyHint: true, mutations get destructiveHint: true
  • Multiple schema sources — SDL files, globs, introspection JSON, inline SDL strings, or live URL introspection
  • Multi-endpoint — combine multiple GraphQL APIs into a single MCP server with prefix-based namespacing
  • Include/exclude filters — cherry-pick which operations become tools
  • ESM only — modern, tree-shakeable, with complete TypeScript types

Quick Start

Library Mode

Add GraphQL tools to an existing MCP server:

import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
import { StreamableHTTPServerTransport } from '@modelcontextprotocol/sdk/server/streamableHttp.js';
import { registerGraphQLTools } from '@graphql2mcp/lib';

const server = new McpServer({ name: 'my-server', version: '1.0.0' });

// Register your own tools alongside GraphQL tools
registerGraphQLTools(server, {
    source: 'schema.graphql',
    endpoint: 'https://api.example.com/graphql'
});

const transport = new StreamableHTTPServerTransport({ sessionIdGenerator: undefined });
await server.connect(transport);

Or use getGraphQLTools for full control over registration:

import { getGraphQLTools } from '@graphql2mcp/lib';

const { tools } = getGraphQLTools({
    source: 'schema.graphql',
    endpoint: 'https://api.example.com/graphql'
});

for (const tool of tools) {
    server.registerTool(
        tool.name,
        {
            title: tool.title,
            description: tool.description,
            inputSchema: tool.inputSchema,
            annotations: tool.annotations
        },
        tool.handler
    );
}

Proxy Mode

Run against a live GraphQL endpoint (introspects the schema automatically):

npx graphql2mcp https://api.example.com/graphql -t http

This starts a Streamable HTTP MCP server on port 3000. For stdio transport (used by Claude Desktop, Cursor), omit the -t http flag:

npx graphql2mcp https://api.example.com/graphql

Or from a local SDL file:

npx graphql2mcp schema.graphql -e https://api.example.com/graphql -t http

Packages

This is a monorepo managed with pnpm workspaces:

Package Description
@graphql2mcp/lib Library for integrating into existing TypeScript MCP servers
@graphql2mcp/core Shared engine — schema loading, tool generation, execution, and MCP server registration
graphql2mcp Standalone CLI proxy — point at a GraphQL endpoint, get an MCP server

How It Works

flowchart TD
    subgraph Input
        A[SDL File] --> L
        B[Inline SDL] --> L
        C[Introspection JSON] --> L
        D[Glob Pattern] --> L
        E[Live URL] --> U
    end

    L[loadSchema] --> S[GraphQLSchema]
    U[loadSchemaFromUrl] --> S

    subgraph "@graphql2mcp/core"
        S --> G[generateTools]
        G --> T[ToolDefinition]
    end

    subgraph MCP Server
        T --> R[registerTool]
        R --> Q[Tool Called]
    end

    Q -->|POST query + variables| EP[GraphQL Endpoint]
    EP -->|JSON response| Q
  1. Load — read a GraphQL schema from an SDL file (or glob of multiple files), introspection JSON, inline SDL string, or live URL introspection
  2. Parse — build a GraphQLSchema object using the graphql library
  3. Generate — walk every Query and Mutation field, mapping arguments to Zod schemas, building field selections, and producing ToolDefinition objects with names, descriptions, annotations, and pre-built query documents
  4. Register — add each tool to an McpServer. When an AI agent calls a tool, the server executes the corresponding GraphQL operation against the endpoint and returns the result as JSON

Runtime Compatibility

Runtime Version Status
Node.js >= 22 Tested
Bun >= 1.2 Tested
Deno >= 2.0 Tested

Development

pnpm install
pnpm build
pnpm test

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选