AI-powered webhook tunnels — inspect, explain and replay every request in real time. Also works for an MCP server or any other HTTP service on your machine.
$ curl -fsSL pipehero.app/install | shWorks with Stripe, GitHub, Shopify webhooks — or a Streamable HTTP MCP server
$ pipehero start myapp --port 3000 tunnel up https://myapp.t.pipehero.app → localhost:3000 inspect http://127.0.0.1:4140 POST /webhooks/stripe 200 24ms POST /webhooks/github 200 18ms GET /health 200 3ms POST /webhooks/stripe 500 61ms
A Streamable HTTP MCP server on localhost is invisible to Claude.ai and ChatGPT — they only run in the cloud, with no way to reach your machine. The same tunnel that already works for webhooks gives it a public HTTPS URL.
pipehero start myserver --port 3000 --long-requests
SSE doesn't stream through the tunnel yet, and requests default to a 30-second timeout — --long-requests opts the tunnel into 120 seconds instead, for tool calls that legitimately run long. Set once, remembered for every future run.
Pipehero ships an MCP server, so your AI agent can list your tunnels, read captured webhooks, and replay them to localhost — no context-switching. Different from the tunnel above: this is Pipehero's own MCP, for debugging your webhook traffic — not exposing a server you built. MCP is an open standard, so it works with whatever you use:
claude mcp add --transport http pipehero https://mcp.pipehero.app/mcp
›“The last webhook to myapp returned 500 — read the payload and my handler and tell me why.”
›“Fix it, then replay the webhook to confirm it works.”
Not just a tunnel — a full inspection and replay workflow built for the way developers actually work.
A stable URL per project. Point Stripe, GitHub or any provider at it and receive real webhooks on your machine.
Every request as it lands — method, path, headers, body, status and latency — in a built-in local panel.
Re-fire any request to localhost without waiting for the provider. Tweak the body first to test edge cases.
Connections, request rate and p50/p90 latency — ngrok-style — in your terminal and the web panel.
Share tunnels and request history across your org, with roles and reserved subdomains. Debug together, not over screenshots.
Verify Stripe / GitHub / Shopify signatures at the edge — every request gets a ✓/✗ badge, checked over the raw body before it hits your code.
Expose a local Streamable HTTP MCP server to Claude.ai, ChatGPT or any cloud-only client — the same tunnel, pointed the other way.
Tunnel your own app's WebSocket connection to localhost — test a real iOS/Android build against your machine over HTTPS, from anywhere. Pro/Team.
Install
curl -fsSL pipehero.app/install | shLog in
pipehero loginStart a tunnel
pipehero start myapp --port 3000Add the public URL to your provider's webhook settings — Stripe, GitHub, Shopify or any HTTP source.
A Streamable HTTP MCP server on localhost is invisible to Claude.ai and ChatGPT. Same tunnel, pointed the other way.
Same tunnel, WebSockets included — point your iOS/Android build at a real HTTPS URL and reach your machine's localhost. Pro/Team.
Start free. Upgrade when you need more.
Everything in Free, plus:
Everything in Pro, plus:
Plans are per workspace — each workspace is billed separately.
Tunnel, inspect, replay. Get back to building.
Open the dashboard