curl-curl-cool-mcp

curl-curl-cool-mcp

An MCP server that gives Claude a full HTTP client — make requests, compare environments, infer TypeScript types from live responses, and work directly with your existing Postman collections and environments.

Category
访问服务器

README

curl-curl-cool-mcp

CI

An MCP server that gives Claude a full HTTP client — make requests, compare environments, infer TypeScript types from live responses, and work directly with your existing Postman collections and environments.

Built with the Model Context Protocol TypeScript SDK, native fetch (no Axios), and a provider pattern that keeps every tool unit-testable without network calls.


Tools

Tool Description
http_request Make an HTTP request and return status, headers, and body. JSON responses are pretty-printed. Supports auth, custom headers, body, and {{variable}} substitution.
diff_environments Fire the same request against multiple environments in parallel and show what differs in status codes and response bodies — field by field.
generate_types Infer TypeScript interface declarations from a JSON string. Nested objects become named interfaces. Chain after http_request to generate types from a live API response.
load_collection Parse an OpenAPI 3.0 spec or Postman Collection v2.1 and list all endpoints. Shows pre-defined headers, body, and collection variables.
load_postman_environment Load a Postman environment file and list its variables, masking secrets. Use the values to resolve {{placeholders}} in collection URLs.
get_response_history List recent requests made in this session with status and timing. Resets when the server restarts.

http_request demo


Installation

Via npm (recommended)

npm install -g curl-curl-cool-mcp

Or use it without installing — npx will fetch and run it on demand (see Claude config below).

From source

git clone https://github.com/dicoy/curl-curl-cool-mcp.git
cd curl-curl-cool-mcp
npm install
npm run build

Add to Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json on macOS:

{
  "mcpServers": {
    "httpclient": {
      "command": "npx",
      "args": ["-y", "curl-curl-cool-mcp"]
    }
  }
}

Add to Claude Code

claude mcp add httpclient -- npx -y curl-curl-cool-mcp

Works with Postman

This is where it gets interesting. If you configure this server alongside Postman's official MCP server, Claude becomes the orchestrator between them — browsing your Postman workspace with one server and executing requests with this one.

Why this matters

Postman's MCP server gives Claude access to your workspace: it can list collections, read request definitions, and fetch environment values directly from the Postman API. But it doesn't execute requests.

This server is the execution layer. Together, they let Claude go from "find the endpoint" to "run it, diff it across environments, and generate TypeScript types for the response" in a single conversation — without you copy-pasting a single URL.

Setup

Add both servers to your Claude config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "httpclient": {
      "command": "npx",
      "args": ["-y", "curl-curl-cool-mcp"]
    },
    "postman": {
      "command": "npx",
      "args": ["-y", "@postman/postman-mcp-server"],
      "env": {
        "POSTMAN_API_KEY": "your-api-key-here"
      }
    }
  }
}

Get your Postman API key from Postman → Account Settings → API Keys.
See Postman's MCP server docs for the full setup guide.

Example workflows

Once both servers are configured, you can give Claude prompts like these:

Explore and execute:

"Find the Create Order endpoint in my Payments collection on Postman, load the staging environment, and make a test request."

Claude calls Postman MCP to find the endpoint definition, then calls load_postman_environment and http_request here to execute it.

Cross-environment diff:

"Run the GET /users endpoint from my API collection against dev and prod and tell me what's different."

Claude resolves the base URLs from two environment files and calls diff_environments, which fires both requests in parallel and shows field-level diffs.

Type generation from live data:

"Hit the /products endpoint on staging and generate TypeScript types for the response."

Claude chains http_request → generate_types — no copy-pasting JSON.

Postman integration demo

Variable substitution

Postman collections use {{variable}} placeholders for environment-specific values. Both http_request and diff_environments resolve these natively, so Claude can pass data from Postman MCP straight through without manual substitution.

Single request with variables:

http_request({
  url: "{{base_url}}/api/{{version}}/users/{{user_id}}",
  variables: {
    base_url: "https://api.example.com",
    version:  "v2",
    user_id:  "42"
  }
})

Multi-environment diff with per-env variable maps:

diff_environments({
  path: "{{base_url}}/api/users",
  environments: {
    dev:  { base_url: "https://api-dev.example.com" },
    prod: { base_url: "https://api.example.com"     }
  }
})

Unresolved placeholders are left intact so Claude can see what's missing rather than silently sending a broken URL.

diff_environments demo


Authentication

http_request and diff_environments both support three auth mechanisms via the auth parameter:

Type Parameters Header produced
bearer token Authorization: Bearer <token>
basic username, password Authorization: Basic <base64>
api-key header (default: X-API-Key), value <header>: <value>

Explicit headers always take precedence over auth, so you can override if needed.


Architecture

src/
├── providers/           # Interfaces + implementations (injectable for tests)
│   ├── http.ts          # IHttpProvider / NodeFetchProvider (native fetch, AbortController)
│   ├── history.ts       # IHistoryProvider / InMemoryHistoryProvider (max 20 entries)
│   ├── collection.ts    # ICollectionProvider / FsCollectionProvider (OpenAPI + Postman)
│   └── postman-environment.ts  # IPostmanEnvironmentProvider / FsPostmanEnvironmentProvider
├── tools/               # One directory per tool: schema.ts + handler.ts + handler.test.ts
├── utils/
│   ├── auth.ts          # resolveAuthHeaders — shared by http_request + diff_environments
│   └── variables.ts     # substituteVariables — {{placeholder}} resolution
└── registry/
    └── tool-registry.ts # Wires providers into MCP tool registrations

Each tool handler is a pure function (input, ...providers) => string. Providers are injected, so tests mock the interface rather than the network or filesystem. 44 unit tests, zero network calls.

Tech: TypeScript 5 (strict + exactOptionalPropertyTypes), MCP SDK, Zod, Biome, Vitest, tsup.

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