How to self-host n8n

Steps

1. Get a Linux server with Docker

   The n8n docs name no minimum memory for a self-hosted server. We would start with 2 GB of memory (our guidance, not an n8n figure); a long workflow that holds big lists in memory needs more. Install Docker with the official script.

   $ curl -fsSL https://get.docker.com | sh

2. Start n8n with a volume

   The data folder /home/node/.n8n holds the SQLite database and the encryption key, so it lives in a volume. GENERIC_TIMEZONE is the instance timezone, important for schedule nodes; the default is America/New_York. Set your own. We changed the official command in two places: -d runs it in the background and --restart unless-stopped brings it back after a reboot.

   $ docker volume create n8n_data
   $ docker run -d --restart unless-stopped --name n8n \
       -p 5678:5678 \
       -e GENERIC_TIMEZONE="Europe/Berlin" \
       -e TZ="Europe/Berlin" \
       -e N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true \
       -v n8n_data:/home/node/.n8n \
       n8nio/n8n

3. Open it and create the first account

   Open port 5678 of the server in a browser. On the first visit n8n asks you to create an account; that account owns the instance.

4. Give it an https address

   Telegram, for one, calls webhooks only over https. Put a reverse proxy with a certificate in front of port 5678: n8n's own guide does it with Traefik in Docker Compose and an A record for your subdomain. Then restart the container with the address. N8N_WEBHOOK_URL (WEBHOOK_URL before n8n 2.30.0) tells n8n its public address; without it n8n builds the URL from N8N_PROTOCOL, N8N_HOST and N8N_PORT, which is wrong behind a reverse proxy. N8N_PROXY_HOPS is the number of reverse proxies n8n runs behind; the default is 0.

   $ docker rm -f n8n
   $ docker run -d --restart unless-stopped --name n8n \
       -p 127.0.0.1:5678:5678 \
       -e N8N_WEBHOOK_URL=https://n8n.example.com/ \
       -e N8N_PROXY_HOPS=1 \
       -e GENERIC_TIMEZONE="Europe/Berlin" \
       -e TZ="Europe/Berlin" \
       -e N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true \
       -v n8n_data:/home/node/.n8n \
       n8nio/n8n

   The data stays in the volume, so you lose nothing. 127.0.0.1 keeps port 5678 closed to the outside: a published port is open on all network interfaces by default; put 127.0.0.1 first in the publish string to keep it local.

5. Build a workflow that starts on a webhook

   Add a Webhook node and set its method to POST. A production webhook is registered only when you publish the workflow; the test URL works while you listen for a test event in the editor. The production path is /webhook/ followed by the path you chose; the test path is /webhook-test/.

6. Update when a new version comes out

   Pull the image, then remove the container and run the same command again. The volume keeps your workflows and credentials.

   $ docker pull n8nio/n8n
   $ docker rm -f n8n

Check that it works

Publish the workflow, then call its production address. A run appears in the Executions tab of n8n.

$ curl -X POST https://n8n.example.com/webhook/hello \
    -H "Content-Type: application/json" -d '{"name":"test"}'

Triggers on a computer that sleeps

The two triggers that start most workflows behave differently once the server sleeps.

The fix is to let the platform keep the clock. Uno has scheduled tasks: a command runs on your computer on a cron schedule, and Uno keeps the history and the log of every run. In short: a scheduled task that comes due wakes a sleeping computer and runs its command. Point that task at a Webhook node on the same computer, and n8n runs the workflow on time.

# every weekday at 9:00 UTC: wake the computer and call the workflow
$ curl -s -H "Authorization: Bearer $UNO_API_KEY" -X POST \
    https://console.uno.place/api/v1/scheduled-tasks \
    -d '{"name":"n8n daily report","box_id":42,"cron_expr":"0 9 * * 1-5","timezone":"UTC",
         "command":"curl -fsS -X POST http://localhost:5678/webhook/daily-report"}'

The same line in the crontab of the computer would not help: a crontab entry inside a sleeping computer does not run and cannot wake it. If a workflow must run when nobody calls it and nothing wakes the computer, use a server that never sleeps.

License and size

Without a license key n8n runs as the free Community edition. Its license, the Sustainable Use License, lets you use the software for your own internal business purposes or for non-commercial or personal use, and share it only free of charge for non-commercial purposes. Offering n8n to others for a fee is outside it; read the license before you build a product on it. The n8n docs warn that self-hosting n8n takes technical knowledge: setting up servers and containers, managing resources, and securing servers and applications.

What this does not do

• A workflow with a Schedule Trigger runs only after you save and publish it, and only while n8n is running. A sleeping computer is not running it.
• Small has 1 GB of memory and Plus has 4 GB. We would start with 2 GB of memory, so n8n belongs on Plus.
• An agent sets this up for you only from Plus: on Small the computer is a bare server for your own agent; Uno Work in the cloud starts with Plus. On Small you run the steps above yourself.

Common problems