Pipehero

io.github.pipehero由 GitHub 账号 pipehero 发布v0.1.0更新于 Nov 2, 2026

Debug webhooks from your AI agent: inspect and replay captured webhooks on localhost.

已验证Streamable HTTP可网页运行云端与基础设施开发者工具安全与监控

概览

AI 生成的概览

Pipehero 把本地 localhost 通过隧道暴露到公网,并让 AI 助手列出隧道、查看与重放捕获的 webhook,以及管理 API 文档。

可以做什么
  • 列出隧道及其在线状态,浏览最近捕获的 webhook 的方法、路径、状态与请求响应体
  • 查看单条捕获请求与响应的完整内容,重放到 localhost,或发送 Stripe、GitHub、Shopify 等服务的模拟测试 webhook
  • 创建、更新、删除并记录 API 端点与集合,添加响应示例,导入或导出 OpenAPI、Postman 与 curl 规格
  • 按环境运行端点,把捕获的真实流量保存为文档,并对比文档与实际流量发现偏差
  • 查看定时 API 监控、待审核的变更提案以及端点最近的运行记录
  • 为单个端点或整个集合创建并查看只读分享链接
适用场景
适合在本地开发 webhook 集成时,让助手查看实际投递内容、解释处理函数失败原因,并在修复后重放验证。也适合把 API 文档放在 Pipehero 的团队,让助手随代码和真实流量同步文档。
运行要求
远程 MCP 端点 mcp.pipehero.app,通过 streamable HTTP 连接;通过浏览器 OAuth 登录,CI 或无头客户端可在 Authorization 头中传 bearer token。也可改用本地 CLI 安装。重放、测试 webhook 和本地环境运行需要隧道 CLI 正在运行;部分功能需要 Pro 或 Team 套餐。
安装前请注意
助手可以重放请求、向本地发送测试 webhook,并可创建、更新或删除 API 端点与分享链接;端点删除无法撤销。分享可能让文档对任何持有链接的人可见。对 staging 或生产环境的运行会从 Pipehero 云端发出真实请求。部分工作区会把 AI 的写入交由同事审核。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Pipehero,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.pipehero%2Fpipehero/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "pipehero": {
      "type": "http",
      "url": "https://mcp.pipehero.app/mcp"
    }
  }
}

README

Pipehero — MCP server + skill

Pipehero tunnels your localhost to a public URL and captures every webhook, so you can inspect, replay, and debug them — including from your AI agent via MCP.

This repo is auto-published from Pipehero's source. It holds the public MCP manifest and the agent skill. Issues/PRs welcome.

Connect the MCP server

Let an agent (Claude Code, Cursor, Copilot, Windsurf, Zed, …) list your tunnels, read captured webhooks, and replay them to localhost.

Remote — recommended (no install, OAuth)

bash
claude mcp add --transport http pipehero https://mcp.pipehero.app/mcp

Your client opens the browser to sign in and approve — no token to paste.

For CI/headless, pass a token from https://pipehero.app/dashboard/settings:

bash
claude mcp add --transport http pipehero https://mcp.pipehero.app/mcp \  --header "Authorization: Bearer <token>"

Local — via the CLI

bash
curl -fsSL https://pipehero.app/install | shpipehero loginclaude mcp add pipehero -- pipehero mcp

Any MCP client works — add it to .cursor/mcp.json, claude_desktop_config.json, etc.:

json
{ "mcpServers": { "pipehero": { "command": "pipehero", "args": ["mcp"] } } }

Tools

ToolWhat it does
list_tunnelsYour tunnels and whether each is online.
list_requests(subdomain)Recent captured webhooks for a tunnel.
get_request(subdomain, id)Full request + response (headers + body).
replay_request(subdomain, id)Replay a webhook to your localhost.
start_tunnel(name, port)Expose a local port on a public URL (local MCP only).
stop_tunnel(name)Stop a tunnel started with start_tunnel (local MCP only).
list_api_collections, get_api_collection(collection_id)Your saved API docs: folders, endpoints, environments (remote MCP, and local from 0.1.14).
get_api_endpoint(collection_id, method + path)One endpoint in full, with its version and response examples.
upsert_api_endpoint(collection_id, method, path, …)Create or update an endpoint by route, so docs follow your code.
add_api_example(collection_id, method + path, name, status, …)Save a response example.
delete_api_endpoint(collection_id, method + path)Remove an endpoint whose route is gone.
import_api_spec(format, content) / export_api_collection(collection_id)OpenAPI, Postman and curl, both ways.
run_api_endpoint(collection_id, method + path, environment)Send it and read the real response.
save_captured_request(collection_id, tunnel, request_id)Turn a real captured request into docs, without credentials.
check_api_drift(collection_id, tunnel)Compare the docs with real traffic (Pro/Team).
list_api_monitors(collection_id)Which endpoints are monitored and which are failing, with the last error (Pro/Team).
list_api_proposalsWhat happened to your proposed changes, on workspaces that review AI changes.
get_api_share_link(collection_id, method + path)The endpoint's shareable link: a read-only page with its docs, examples and latest run.
share_api_endpoint(collection_id, method + path, …)Create that link (workspace-only by default), open it to anyone with the link, or invite people by email.
share_api_collection(collection_id, …)The same for a whole collection: every endpoint, read-only, never its environments.

Because your agent has both the captured webhook and your codebase, it can explain why a handler failed — then fix it and replay to confirm.

Skill

Install the skill so your agent knows when and how to use Pipehero:

bash
pipehero skill install

or copy SKILL.md into ~/.claude/skills/pipehero/SKILL.md.

License

MIT

来源:README.md,提交 6473cb0

工具

25

list_tunnels

List tunnels

List the user's Pipehero tunnels and whether each is currently online.

无需输入参数。

list_requests

List captured requests

List recent webhooks/requests captured for a tunnel. Returns id, method, path, status and bodies.

输入参数subdomain必填

get_request

Get a captured request

Get the full request and response (headers + body) of one captured webhook by id.

输入参数id必填subdomain必填

replay_request

Replay a webhook

Replay a captured webhook to the user's localhost (the tunnel CLI must be running).

输入参数id必填subdomain必填

send_test_webhook

Send a test webhook

Send a generated test webhook to the user's localhost (the tunnel CLI must be running). Provide a `provider` (stripe, github, shopify, clerk, slack, resend, paddle, polar, lemonsqueezy) and optional `event` for a realistic sample payload, or pass a custom JSON `body`. Pro plan required.

输入参数bodyeventprovider必填subdomain必填

set_long_requests

Set extended request timeout

Opt a tunnel into a 120-second ingress timeout instead of the 30-second default — for tool calls that legitimately run long, e.g. an MCP server exposed through the tunnel. Persists on the tunnel; pass enabled: false to revert to the 30-second default.

输入参数enabled必填subdomain必填

set_ws_passthrough

Set WebSocket passthrough

Opt a tunnel into WebSocket passthrough, so an application-level WebSocket connection (e.g. a mobile/web app's own realtime feature) can be relayed to your localhost. Pro/Team plan required to enable; pass enabled: false to disable on any plan.

输入参数enabled必填subdomain必填

list_api_collections

List API collections

List the workspace's API collections (saved endpoints with examples and environments). Start here to find a collection_id.

无需输入参数。

get_api_collection

Get an API collection

Get one collection's folders, endpoints (id, method, path, name, version, example count) and environments (with their variables).

输入参数collection_id必填

get_api_endpoint

Get an API endpoint

Get one endpoint in full: description, parameters, headers, request body, saved response examples and its current `version`. Identify it by endpoint_id or by method + path.

输入参数pathmethodendpoint_idcollection_id必填

create_api_collection

Create an API collection

Create an empty API collection. Prefer import_api_spec when there is an OpenAPI file, and upsert_api_endpoint to fill it from the code.

输入参数name必填description

upsert_api_endpoint

Create or update an API endpoint

Document an endpoint: creates it, or updates the existing one for the same method + path. Use it whenever you add or change a route in the code, so the docs stay true. Only the fields you send change. `path` may use :id, {id} or {{id}} for variables. `folder` is created by name if missing. Send `version` (from get_api_endpoint) to avoid overwriting a teammate's edit; a stale version is rejected.

输入参数bodykindnamepath必填foldermethod必填paramsheadersversiondescriptioncollection_id必填

add_api_example

Add a response example

Save a response example on an endpoint (for instance the 402 your handler returns for a declined card). Replaces the example with the same name, so it is safe to repeat.

输入参数bodyname必填pathmethodstatus必填headersendpoint_idstatus_textcollection_id必填

delete_api_endpoint

Delete an API endpoint

Delete an endpoint and its examples. Use when the route no longer exists in the code. Cannot be undone.

输入参数pathmethodendpoint_idcollection_id必填

import_api_spec

Import an API spec

Import an OpenAPI 3.x document (as JSON), a Postman v2.x collection (JSON) or a curl command into a new collection, or into `collection_id`. All or nothing, and plan limits are checked first. Credentials are never imported (they become variables). Convert YAML to JSON before sending.

输入参数nameformat必填content必填collection_id

export_api_collection

Export an API collection

Export a collection as an OpenAPI 3.1 or Postman 2.1 document. Handy to read the whole API at once.

输入参数formatcollection_id必填

run_api_endpoint

Run an API endpoint

Send an endpoint against an environment and return the real response (status, headers, body up to 20 KB). A `local` environment goes through the user's tunnel to their localhost (the CLI must be running); staging/prod are sent from Pipehero's cloud (Pro/Team) to public addresses only. Sending to a `prod` environment with a write method needs confirm_write_to_prod: true, which you should only pass after the user agrees. Use it to check that the code you wrote behaves as documented.

输入参数pathmethodvariablesendpoint_idenvironment必填collection_id必填confirm_write_to_prod

save_captured_request

Save a captured request as docs

Document a real request captured by a tunnel (get its id from list_requests). If an endpoint already covers the route, the response is added to it as an example; otherwise an endpoint is created from the request. Credentials, cookies and signatures are replaced by {{variables}} and only JSON bodies are kept. A response already saved as an example is not duplicated. Use it to document webhooks and calls you have only seen in traffic.

输入参数pathfoldermethodtunnel必填request_id必填endpoint_idcollection_id必填

list_api_monitors

List API monitors

See which endpoints of a collection are checked on a schedule (Pro/Team monitors) and how they are doing: state (healthy, failing, blip, paused, waiting), the last check with its status, latency and error, and since when it has been failing. Use it to find out what is broken in staging or production, then read the endpoint and its code. Read-only: people decide what to monitor.

输入参数collection_id必填

list_api_proposals

List proposed API changes

See what happened to the changes you proposed. Some workspaces review AI changes: your write tools then answer `pending_review` and change nothing until a teammate approves. This lists the latest proposals with their status (pending, approved, rejected with the reviewer's reason, or failed with why) so you can tell whether a change went through.

无需输入参数。

check_api_drift

Check API docs against real traffic

Compare a collection's docs with the real traffic a tunnel captured (Pro/Team). Returns routes nobody documented, status codes with no example, response or request fields missing from every example (or with another type), undocumented query parameters, and documented fields that never appear. Each finding includes `request_id` of the newest request that shows it: pass it to save_captured_request to add it to the docs, or fix the docs with upsert_api_endpoint and add_api_example. Use it to find out what drifted after a code change.

输入参数limittunnel必填collection_id必填

share_api_endpoint

Share an endpoint

Create an endpoint's shareable link and return it. With no other arguments the link stays restricted: workspace members can open it (they sign in), which is enough for a PR or an internal message. Only when the user asks to share it outside the workspace: set `link_access` to `viewer` (anyone with the link can read) or `commenter` (Team plan; commenting asks for sign-in), or `invite` people by email (they get an email and sign in with that address). `restricted` turns the public link off; the URL stays the same. The page is read-only and always hides credentials and secrets.

输入参数pathinvitemethodendpoint_idlink_accesscollection_id必填

share_api_collection

Share a collection

Share a whole collection: every endpoint in it (including ones added later), with its examples and latest run, read-only and with secrets hidden. Environments are never shown and it can't be exported. Returns the collection's link, which never changes. With no other arguments it stays restricted to the workspace; set `link_access` or `invite` people only when the user asks to share it outside the team. get_api_collection returns the link as `share` once it exists.

输入参数invitelink_accesscollection_id必填

list_api_runs

List recent runs

Recent runs of an endpoint (newest first, default 5): who/what was sent and what came back. Use to see what happened the last time it was run.

输入参数pathlimitmethodendpoint_idcollection_id必填

版本历史

1
  1. v0.1.0最新Sep 16, 2026

评分

已满足适用检查项的 90%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分

    请在仓库或注册中心条目中声明许可证。