
MCP Swagger Server
Automatically converts REST APIs with Swagger/OpenAPI documentation into MCP tools, making any documented API accessible to MCP-compatible clients. Supports authentication, SSL handling, and flexible configuration for seamless API integration.
README
MCP Swagger Server
An MCP (Model Context Protocol) server that automatically converts REST APIs with Swagger/OpenAPI documentation into MCP tools, making them accessible to MCP-compatible clients.
Features
- 🔄 Automatic Tool Generation: Converts Swagger/OpenAPI endpoints to MCP tools
- 🔒 SSL Certificate Handling: Option to ignore SSL certificate errors for internal APIs
- 🏷️ Custom Tool Prefixes: Organize tools with custom prefixes for better organization
- 📡 Stdio Transport: Uses stdio format as the default transport mechanism
- 🌐 Flexible Input: Supports both URL and local file swagger documentation
- 🔧 Parameter Support: Handles path, query, and body parameters
- 📝 Type Mapping: Maps Swagger types to JSON Schema types for proper validation
Installation
Global Installation
npm install -g mcp-swagger
From Source
git clone https://github.com/HainanZhao/mcp-swagger.git
cd mcp-swagger
npm install
npm run build
Usage
Command Line Options
mcp-swagger [options]
Options:
-u, --swagger-url <url> URL to swagger documentation
-f, --swagger-file <file> Path to local swagger file
-p, --tool-prefix <prefix> Custom prefix for generated tools
-b, --base-url <url> Override base URL for API calls
--ignore-ssl Ignore SSL certificate errors
-a, --auth-header <header> Authentication header (e.g., "Bearer token")
-h, --help Display help information
-V, --version Display version number
Examples
Load from URL with custom prefix
mcp-swagger --swagger-url https://api.example.com/swagger.json --tool-prefix example --ignore-ssl
Load from local file
mcp-swagger --swagger-file ./api-docs.json --tool-prefix local-api
With authentication
mcp-swagger --swagger-url https://api.example.com/swagger.json --auth-header "Bearer your-token-here"
Override base URL
mcp-swagger --swagger-file ./swagger.json --base-url https://staging.api.com --tool-prefix staging
Configuration
The server can be configured through command-line arguments or environment variables:
CLI Option | Environment Variable | Description |
---|---|---|
--swagger-url |
SWAGGER_URL |
URL to swagger documentation |
--swagger-file |
SWAGGER_FILE |
Path to local swagger file |
--tool-prefix |
TOOL_PREFIX |
Custom prefix for generated tools |
--base-url |
BASE_URL |
Override base URL for API calls |
--ignore-ssl |
IGNORE_SSL=true |
Ignore SSL certificate errors |
--auth-header |
AUTH_HEADER |
Authentication header |
MCP Integration
Add to your Agent configuration (e.g. claude_desktop_config.json
or ~/.gemini/settings.json
):
{
"mcpServers": {
"swagger-api": {
"command": "mcp-swagger",
"args": [
"--swagger-url", "https://example.com/swagger.json",
"--tool-prefix", "example",
"--ignore-ssl"
]
}
}
}
Other MCP Clients
The server uses the standard MCP stdio transport, so it should work with any MCP-compatible client. Start the server and connect via stdin/stdout.
Generated Tools
Tool Naming Convention
Tools are named using the following pattern:
{prefix}_{method}_{path_segments}
- Path parameters are converted to
by_{parameter_name}
Examples:
GET /v1/users/
→myapi_get_v1_users
GET /v1/users/{id}
→myapi_get_v1_users_by_id
POST /v1/users/
→myapi_post_v1_users
Parameter Mapping
- Path parameters: Included in the URL path
- Query parameters: Added as URL query string
- Body parameters: Sent as JSON request body
- Header parameters: Added to request headers
Type Mapping
Swagger Type | JSON Schema Type |
---|---|
string |
string |
integer |
number |
boolean |
boolean |
array |
array |
object |
object |
Sample Swagger Document
The server has been tested with the following sample swagger document structure:
{
"swagger": "2.0",
"info": {
"title": "API Documentation",
"version": "1.0"
},
"host": "api.example.com",
"basePath": "/api",
"paths": {
"/v1/hosts/": {
"get": {
"summary": "Get a list of hosts",
"parameters": [
{
"name": "filter_by",
"in": "query",
"type": "string",
"description": "Filter criteria"
}
]
}
},
"/v1/hosts/{name}": {
"get": {
"summary": "Get host by name",
"parameters": [
{
"name": "name",
"in": "path",
"required": true,
"type": "string"
}
]
}
}
}
}
This would generate tools like:
example_get_v1_hosts
- List hosts with optional filteringexample_get_v1_hosts_by_name
- Get specific host by name
Error Handling
The server includes comprehensive error handling:
- SSL Certificate Errors: Can be ignored with
--ignore-ssl
flag - Network Errors: Returned as error responses with details
- Invalid Swagger: Validation errors are reported during startup
- Missing Parameters: Parameter validation based on swagger schema
- HTTP Errors: API response errors are captured and returned
Development
Building
npm run build
Testing
# Test with sample swagger file
npm run test
Linting
npm run lint
Troubleshooting
Common Issues
-
SSL Certificate Errors
- Use
--ignore-ssl
flag for internal APIs with self-signed certificates
- Use
-
Tool Name Conflicts
- Use
--tool-prefix
to add unique prefixes to avoid naming conflicts
- Use
-
Base URL Issues
- Use
--base-url
to override the base URL from swagger documentation
- Use
-
Authentication Failures
- Provide proper authentication header with
--auth-header
- Provide proper authentication header with
Debug Mode
The server logs important information to stderr:
- Swagger document loading status
- Number of tools generated
- Tool generation details
License
MIT License - see LICENSE file for details.
Contributing
- Fork the repository
- Create a feature branch
- Make your changes
- Add tests if applicable
- Submit a pull request
Support
For issues and questions:
- Create an issue on GitHub
- Check the troubleshooting section above
- Review the sample configurations
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。