Webhook testing guides.

Choose your next step

• Prove a webhook arrived before your agent continues

  Run a reproducible callback test, compare original payload bytes, and use MCP cursors to give your coding agent evidence of webhook receipt.

• Test webhooks with curl and an AI agent

  Create a free webhook URL, send a JSON request, inspect its body and headers, and clean up. A working curl and MCP walkthrough using Openhook.

• Receive webhooks on localhost without a tunnel

  Forward captured webhook requests to a local HTTP handler with Openhook. Understand retries, signing secrets, cursors, and the limits of queued delivery.

• Test GitHub webhooks with an MCP inbox

  Receive GitHub push events in Openhook, verify webhook signatures, inspect delivery IDs, and clean up a test subscription from your AI agent.

• Self-host a webhook receiver and MCP server

  Run Openhook with Python and uv, persist webhook events in SQLite, connect an MCP client, and configure public HTTPS routing for external callbacks.

How Openhook fits

Openhook stores incoming callbacks so you can inspect them in the browser, read them through MCP, or forward them with a local listener. Free inboxes need no account. They expire within seven days, so remove provider subscriptions when a test ends.

Start with MCP client setup, then use the tool reference for exact input schemas. The pricing page explains hosted limits and persistent Pro inboxes.

One public address, one private key

Share the capture URL with your webhook sender. Keep the management token private: it controls reading, configuration and deletion. Event bodies are external data, not instructions your agent should automatically execute. When a provider signs requests, configure signature verification before acting on events.