Hello everyone,
I have a self-hosted n8n 2.13.2 instance with Docker Compose, here is the docker-compose.yml:
services:
n8n:
image: docker.n8n.io/n8nio/n8n
restart: unless-stopped
environment:
- N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true
- N8N_RESTRICT_FILE_ACCESS_TO=‘/files’
- N8N_HOST=n8n.myurl.com
- N8N_PORT=5678
- N8N_PROTOCOL=https
- NODE_ENV=production
- N8N_PROXY_HOPS=1
- WEBHOOK_URL=https://n8n.myurl.com/
- GENERIC_TIMEZONE=${GENERIC_TIMEZONE}
- TZ=${GENERIC_TIMEZONE}
volumes:- ./n8n_data:/home/node/.n8n
- ./local_files:/files
networks:- traefik
- backend
networks:
traefik:
external: true
backend:
name: mynw_backend
I would like to use the n8n connector offered on Claude.ai, but I’m running into a problem after granting the Claude.ai connector permission to use my instance, with the following message:
Authorization with the MCP server failed. You can check your credentials and permissions. If this persists, share this reference with support: “ofid_1728793076d22151”
To allow Claude.ai to access my instance, I set up a Cloudflare tunnel with, for now, a very open configuration in terms of access rules:
Policies
In the logs of my n8n container, I can clearly see the connection request and its acceptance:
n8n-1 | Consent approved
n8n-1 | Authorization code exchanged for tokens
And the request shows up in the connected clients on the n8n interface:
But on the Claude.ai side, the connector is still not… connected.
I have read the n8n MCP documentation on instance-level mode several times, but I haven’t found any specific information that could help me identify the cause of the problem.
Do you have any leads?
Thanks in advance.
Information on your n8n setup
- n8n version: 2.13.2
- Database : SQLite
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via : Docker
- Operating system: Linux