mcpdirectory.dev
9,129 servers · updated 23 Sept

Webhooks MCP server

community Active last commit 33 d
{
  "mcpServers": {
    "webhook-studio": {
      "type": "http",
      "url": "https://webhook-studio.com/mcp"
    }
  }
}
file .mcp.json (in your project root)version 0.1.1

What it does

Webhook Studio is an MCP server that provisions temporary webhook endpoints, captures incoming events, replays them locally, forwards them to backends, and generates TypeScript/Zod types from real traffic. It is for developers and AI agents integrating webhook providers without signup or keys.

  • Debug webhook integrations by capturing full request headers, bodies, and signature verification results in a live inbox.
  • Forward webhook events to your application with conditional routing and automatic re-delivery on failures.
  • Generate accurate TypeScript, Zod, or JSON-Schema types by analyzing the actual webhook payloads your account receives.

Tools this server exposes

ToolDoes
create_bucketProvision a new webhook endpoint.
list_bucketsList all webhook endpoints.
get_bucketRetrieve details of a specific webhook endpoint.
delete_bucketDelete a webhook endpoint.
latest_eventGet the most recent event received on an endpoint.
list_eventsList events received on an endpoint.
get_eventRetrieve details of a specific event.
wait_for_eventBlock until an event is received on an endpoint.
replay_eventReplay a captured event to a local endpoint.
forward_bucketConfigure forwarding of events from an endpoint to a backend URL.
list_deliveriesList forwarded event deliveries and their results.
configure_verificationEnable HMAC signature verification with a provider's secret.
list_schemasList learned schemas from captured events.
get_schemaRetrieve the schema of captured events.
diff_schemaCompare schemas between two versions to identify changes.
generate_typesGenerate TypeScript, Zod, or JSON-Schema types from captured payloads.
find_correlation_keysIdentify fields that correlate events across webhook deliveries.

Before you enable it

An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.