hookray-mcp

hookray-mcp

Enables AI agents to create disposable webhook URLs, capture incoming HTTP requests, inspect headers and bodies, and replay them against local or remote endpoints, streamlining the webhook handler development loop.

Category
访问服务器

README

hookray-mcp

Webhook testing for AI agents — capture, inspect, replay HTTP requests from Claude Code, Cursor, Claude Desktop, or any MCP-compatible client.

npm version License: MIT

hookray-mcp is the Model Context Protocol server for HookRay. Drop it into your AI agent's config and the agent can spin up disposable webhook URLs, watch incoming requests live, and replay captured payloads against your local handler — without leaving the chat.

Why this exists

AI coding agents increasingly write webhook handlers (Stripe, GitHub, Shopify, Twilio, …). The agent loop today is:

  1. Generate handler code
  2. Manually test it by triggering a real event
  3. Manually copy/paste the captured payload back to the agent
  4. Agent fixes the bug
  5. Repeat from step 2

With hookray-mcp installed, the agent does steps 2–4 itself:

You: "Build me a Stripe webhook handler at /api/stripe."
Agent: [writes code, then calls hookray-mcp tools]
       → create_webhook_inbox
       → wait_for_request   (you trigger an event in Stripe)
       → inspect_request
       → replay_request to localhost:3000/api/stripe
       → "Returned 500. The bug is in the signature check at line 14."

Install

# Anywhere npx works
npx -y hookray-mcp@latest

# Or globally
npm install -g hookray-mcp

Node.js 20+ required.

Configure your AI client

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "hookray": {
      "command": "npx",
      "args": ["-y", "hookray-mcp@latest"]
    }
  }
}

Claude Code

claude mcp add hookray -- npx -y hookray-mcp@latest

Cursor

Add to .cursor/mcp.json in your project (or the global Cursor MCP config):

{
  "mcpServers": {
    "hookray": {
      "command": "npx",
      "args": ["-y", "hookray-mcp@latest"]
    }
  }
}

With a HookRay API key (for Pro features)

If you have a HookRay account and want persistent inboxes, pass HOOKRAY_API_KEY:

{
  "mcpServers": {
    "hookray": {
      "command": "npx",
      "args": ["-y", "hookray-mcp@latest"],
      "env": {
        "HOOKRAY_API_KEY": "hkr_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
      }
    }
  }
}

Get a key at hookray.com/app/settings/api-keys.

Tools

Tool Description Auth
create_webhook_inbox Generate a fresh https://hookray.com/api/hook/xxxx URL that captures any HTTP request. Returns the URL + inbox id. Anonymous OK. persistent: true needs Pro.
list_requests List recent requests captured by an inbox (with filters: method, since, search, limit). Anonymous OK.
inspect_request Get the full method + headers + body + query params of a single captured request. Anonymous OK.
wait_for_request Block (up to 60 s) until a new request arrives. Critical for AI agent dev loops. Anonymous OK.
replay_request Forward a captured request to any URL — including localhost. Runs client-side from this package so private IPs work. Anonymous OK.

Environment variables

Variable Default Notes
HOOKRAY_BASE_URL https://hookray.com Override for self-hosting or local dev.
HOOKRAY_API_KEY (unset) Required for persistent inboxes. Get one at hookray.com/app/settings/api-keys.

Example: Stripe webhook handler dev loop

> "Add a Stripe webhook handler at /api/stripe/webhook in this Next.js app."
[agent writes app/api/stripe/webhook/route.ts]

> "Now test it."
[agent calls hookray-mcp]
  → create_webhook_inbox        → https://hookray.com/api/hook/abc12345
  → "Paste this URL into Stripe → Add endpoint, then trigger an event."

[you trigger a test event from Stripe dashboard]

  → wait_for_request(inbox_id)  → {method: POST, content_type: application/json, ...}
  → inspect_request(...)        → full headers (incl. stripe-signature) + body
  → replay_request(destination_url=http://localhost:3000/api/stripe/webhook)
                                → {status_code: 400, body: "No signatures..."}
  → "Your handler is reading req.json() before constructEvent. Switch to
     req.text() — Stripe's signature check needs the raw body."
[agent edits the handler]
  → replay_request(...)         → {status_code: 200}

Free vs Pro

Capability Free HookRay Pro
Anonymous inboxes
Capture, list, inspect, wait
Replay to any URL (incl. localhost)
Inbox lifetime 7 days 10 years
Multiple inboxes 1 active up to 10
Request history 100 / month 10,000 / month

Pro is $9 / month at hookray.com/pricing.

Limitations & non-goals

  • stdio transport only for v0.1. SSE / HTTP transports may come later.
  • No automatic agent triggering of real Stripe events — you still need to fire the event yourself from the provider dashboard or CLI. (Provider-specific sample-payload tools are on the roadmap.)
  • No WebSocket realtimewait_for_request long-polls (1 s tick, 60 s max). Good enough for dev loops; not designed for production traffic.
  • SSRF safety on replay_request is your responsibility (the package runs on your machine, in your network).

Roadmap

  • [ ] Provider-specific template tools (send_stripe_test_event, send_github_test_event, …)
  • [ ] SSE transport for hosted MCP gateways
  • [ ] Streamed responses for long-running replays
  • [ ] First-class self-host docs (point at your own HookRay instance via HOOKRAY_BASE_URL)

Development

git clone https://github.com/ShotaTanikawa/hookray-mcp.git
cd hookray-mcp
pnpm install
pnpm build
pnpm test

License

MIT © 2026 Shota Tanikawa. See LICENSE.


Made by @ShotaTanikawa · HookRay · Feedback welcome — open an issue.

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