Contex
Converts any OpenAPI 3.x specification into a fully functional MCP server with OAuth 2.1 support and a purely functional architecture.
README
Contex - OpenAPI to Model Context Protocol Server
A generic TypeScript library for converting OpenAPI specifications into MCP (Model Context Protocol) servers. Built with a purely functional architecture using Ramda.
Features
- Generic OpenAPI Conversion: Convert any OpenAPI 3.x specification into a fully functional MCP server
- OAuth 2.1 Support: Built-in authentication with PKCE support
- Purely Functional: Composable pure functions using Ramda and
R.pipe() - Type-Safe: Full TypeScript support with Result monad for error handling
- Extensible: Easy to add custom tools and resources
- Well-Tested: 100% unit test coverage target
Installation
npm install contex zod
Quick Start
import { createMcpServer, isOk } from 'contex';
const result = await createMcpServer({
name: 'My API Server',
version: '1.0.0',
openApiSpec: './openapi.yaml',
baseUrl: 'https://api.example.com',
});
if (isOk(result)) {
await result.value.start();
console.log('MCP server running on port 3000');
}
With OAuth Authentication
import { createMcpServer, isOk } from 'contex';
const result = await createMcpServer({
name: 'Secure API Server',
version: '1.0.0',
openApiSpec: './openapi.yaml',
baseUrl: 'https://api.example.com',
auth: {
type: 'oauth2',
clientId: process.env.CLIENT_ID,
clientSecret: process.env.CLIENT_SECRET,
authorizationUrl: 'https://auth.example.com/authorize',
tokenUrl: 'https://auth.example.com/token',
scopes: ['read', 'write'],
pkce: true, // Enabled by default
},
server: {
port: 3000,
transport: 'streamable-http',
},
});
if (isOk(result)) {
await result.value.start();
}
API Overview
Main Functions
createMcpServer(config)- Create MCP server from OpenAPI speccreateCustomMcpServer(name, version, tools, resources)- Create server with custom tools
Result Type
The library uses a Result<T, E> monad for error handling:
import { isOk, isErr, fold, formatError } from 'contex';
const result = await createMcpServer(config);
// Pattern matching with type guards
if (isOk(result)) {
const server = result.value;
} else {
console.error(formatError(result.error));
}
// Or use fold
const server = fold(
(error) => { console.error(error); return null; },
(server) => server
)(result);
Tool and Resource Generation
import {
parseOpenApiSpec,
extractOperations,
buildToolsPipeline,
buildResourcesPipeline,
} from 'contex';
// Parse OpenAPI spec
const specResult = await parseOpenApiSpec('./openapi.yaml');
if (isOk(specResult)) {
const spec = specResult.value;
// Extract operations
const operations = extractOperations(spec);
// Build tools
const tools = buildToolsPipeline('https://api.example.com')(operations);
// Build resources
const resources = buildResourcesPipeline(spec);
}
Architecture
See docs/ARCHITECTURE.md for details.
Documentation
- API Reference - Complete API documentation
- Architecture Guide - Design principles and patterns
- Examples - Usage examples and patterns
Examples
See the examples/ directory for complete examples:
basic-server.ts- Basic MCP server from OpenAPIwith-oauth.ts- Server with OAuth authenticationcustom-tools.ts- Server with custom tool definitions
Existing Alternatives
Several similar libraries exist in the ecosystem:
| Library | Description |
|---|---|
| openapi-mcpserver-generator | Generates MCP server code from OpenAPI specs |
| openapi-mcp-generator | TypeScript tool for OpenAPI → MCP conversion |
| openapi-to-mcp-converter | Automatic OpenAPI to MCP Server conversion |
| openapi-mcp-server | Bridge between OpenAPI and MCP for Claude Desktop |
| api-to-mcp | Turn any API into MCP tools |
| Stainless | Commercial platform for MCP server generation |
| Speakeasy | Commercial tool for generating MCP servers |
How This Library Differs
| Feature | Existing Tools | This Library |
|---|---|---|
| Architecture | Mostly imperative/OOP | Purely functional |
| Error Handling | Exceptions | Result monad |
| OAuth Support | Limited/varies | Full OAuth 2.1 + PKCE |
| Runtime | Code generation | Runtime conversion |
| Composability | Low | High (R.pipe) |
Development
# Install dependencies
npm install
# Run tests
npm test
# Type check
npm run typecheck
# Build
npm run build
License
MIT
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。