Discover and register MCP servers via webhooks with HTTP discovery and AI enrichment

Quick overview

This workflow exposes a webhook-based MCP server registry API that discovers MCP servers from a seed list, enriches tool/resource metadata via an HTTP AI enrichment endpoint, upserts entries to a central catalog, performs health checks, logs an audit event, and returns a structured discovery, search, registration, or health report.

How it works

1. Receives a POST request on a webhook endpoint with an action such as discover, register, search, health, or list.
2. Validates the request payload, builds a request context (requestId, query, serverUrl, tools/resources), and selects the downstream endpoint URLs (optionally simulating discovery status codes).
3. Calls an HTTP discovery endpoint to probe a configured seed list of MCP server URLs and collect discovery results.
4. Sends the discovery/context data to an HTTP AI enrichment endpoint to generate improved descriptions, categories, tags, and capability flags.
5. Calls an HTTP registry endpoint to register or update the server, tools, resources, and metadata in a central catalog.
6. Calls an HTTP health endpoint to check server status/latency and then calls an HTTP logging endpoint to write an audit trail entry.
7. Builds an action-specific response (catalog list, registration result, search hits, or health summary) and returns it via the webhook response.

Setup

1. Configure the webhook path in n8n and copy the production webhook URL into any client or system that will call this registry API.
2. Replace the placeholder httpbin.org URLs with your real discovery, AI enrichment, registry, health, and audit logging service endpoints.
3. Update the seed MCP server list (and any default tags) in the workflow’s configuration code to match your environment.
4. If your HTTP endpoints require authentication, add the required headers/credentials to the HTTP Request steps.