Two ways to run it
The quickest is the n8n template in App Hosting. It deploys n8n without a server to patch. If you want root, other containers on the same host, or your own backup routine, use a VPS.
Docker on a VPS
- Create the server
Use the Docker stack image or Ubuntu 24.04, with a hostname pointing at the public IPv4.
- Write a Compose file
Run n8n and a database in Compose, and keep the n8n data on a named volume so it survives restarts.
- Put HTTPS in front
Use a reverse proxy such as Caddy or Nginx to terminate TLS on your hostname, and set the n8n webhook URL to match.
- Set an encryption key
Set the n8n encryption key yourself and store it outside the server, since losing it makes saved credentials unreadable.
- Snapshot and back up
Take the free snapshot and also export workflows and back up the data volume regularly.
Use our own AI endpoint from n8n
The OpenAI node in n8n can use the BookMyHost AI base URL, api.bookmyhost.com/v1. See AI API for models and how to get a key.
Sizing
n8n is light for a small number of workflows. Size by the number of concurrent executions and any heavy nodes, and start on hourly billing, with a 1 hour minimum, until you know the load. Servers start at ₹287 a month.


