Deploy N8N

Deploy and Host n8n on Railway

n8n is a source-available workflow automation platform where you wire triggers, apps, databases, and AI models together on a visual canvas. It ships 400+ native integrations plus a generic HTTP Request node. Deploy n8n on Railway and you get the editor, the execution tier, and the code sandbox as three separate services instead of one overloaded container.

Self-host n8n on Railway with five services wired together at deploy time. n8n serves the editor, REST API, and webhook endpoints from ghcr.io/n8n-io/n8n:2.32.6 on port 5678, and it is the only service with a public domain. Postgres stores workflows, encrypted credentials, execution history, and binary data. Redis is the job queue: because EXECUTIONS_MODE=queue is set, the main process never runs a workflow itself, it enqueues one. n8n-worker runs the same image with the worker command and pulls jobs off Redis. n8n-runner runs ghcr.io/n8n-io/runners:2.32.6-distroless and evaluates Code-node JavaScript and Python inside its own container, attached to the worker's task broker on port 5679. The full request path is Railway edge → n8n:5678 → Redis → worker → runner.

Getting Started with n8n on Railway

Open the generated Railway domain. Self-hosted n8n has no default credentials: the first visit shows the owner-account setup form, and whoever submits it becomes the instance owner, so reach the URL yourself before sharing it. n8n then offers a free community licence key by email that unlocks folders and richer execution debugging; skipping it costs you nothing.

From Overview, click Create Workflow, add a Manual Trigger, add a Code node returning [{ json: { ok: true } }], and press Execute workflow — that one run exercises the entire stack, because the request travels main → Redis → worker → the runner container that actually evaluates your JavaScript.

Open Executions and confirm the run is listed as finished; a run stuck in waiting points at the worker or the Redis connection rather than the editor.

About Hosting n8n

• Visual editor with 400+ integrations — SaaS apps, databases, message queues, and an HTTP Request node for anything else.
• Code nodes in JavaScript and Python — full expression language, npm-style community packages, and per-item or per-run execution modes.
• AI agent building blocks — 70+ AI nodes covering chat models, tools, memory, embeddings, and vector stores for RAG workflows.
• Production triggers — webhooks, cron schedules, polling, and queue consumers, with retries and error workflows.
• Encrypted credential vault — secrets are encrypted at rest in Postgres and never rendered back into the editor.
• Full execution history — inspect, retry, and prune past runs from the UI or automatically.

Why Deploy n8n on Railway

Railway handles the infrastructure work around a five-service n8n stack.

• Managed Postgres and Redis, provisioned with private networking already wired up.
• Free HTTPS domain with automatic TLS, so webhook URLs work immediately.
• Scale throughput by raising worker replicas instead of resizing a VPS.
• Usage-based pricing — pay for the CPU and RAM you actually burn.
• Per-service logs, metrics, variables, and one-click rollbacks.

Common Use Cases for Self-Hosted n8n

• AI agents and RAG assistants. Chain a chat model, a vector store, and your internal APIs as tools so an agent can answer from company documents and take actions, with prompts and embeddings never leaving your infrastructure.
• Scheduled data movement. Cron-triggered syncs between Postgres, spreadsheets, warehouses, and SaaS tools, plus nightly reports, backups, and stale-record cleanups.
• Internal API orchestration. Expose a webhook as a lightweight internal endpoint that fans out to several microservices, normalises their responses, and returns one JSON payload — useful for support tooling and ops runbooks.

Dependencies for n8n

• n8n — ghcr.io/n8n-io/n8n, the official image, current
• Postgres — Railway managed Postgres 13+ with its own volume at /var/lib/postgresql/data. Queue mode does not support SQLite, so Postgres is mandatory here.
• Redis — Railway managed Redis with a volume at /data, acting as the BullMQ job broker for queue mode.

Deployment Dependencies

• Images: ghcr.io/n8n-io/n8n and ghcr.io/n8n-io/runners (also published as docker.n8n.io/n8nio/n8n).

Hardware Requirements for Self-Hosting n8n

n8n is memory-hungry rather than CPU-hungry: a 4-core box with 8 GB of RAM outperforms an 8-core box with 4 GB.

Self-Hosting n8n with Docker

For a quick single-container evaluation with SQLite, the Docker command is:

docker run -d --name n8n -p 5678:5678 \
  -v n8n_data:/home/node/.n8n \
  -e N8N_ENCRYPTION_KEY="$(openssl rand -hex 32)" \
  -e N8N_HOST=localhost -e N8N_PORT=5678 \
  ghcr.io/n8n-io/n8n:2.32.6

n8n vs Zapier vs Make

How Much Does n8n Cost to Self-Host?

The n8n Community Edition is free with no execution cap and no feature metering — self-host it and you pay zero licence fees. n8n Cloud starts around €20/month for 2,500 executions, and paid self-hosted Business and Enterprise tiers exist for SSO, Git-based version control, and multi-project governance. On Railway your cost is infrastructure only: five services (main, worker, runner, Postgres, Redis) billed on usage-based pricing for CPU, memory, egress, and volumes.

One thing to note about the licence: the Sustainable Use License permits internal commercial use freely but restricts reselling n8n itself as a hosted service to third parties.

FAQ

What is n8n?

n8n is a source-available workflow automation and AI agent platform with a visual node editor, 400+ integrations, JavaScript and Python code nodes, webhook and cron triggers, an encrypted credential store, and full execution history. It is licensed under the Sustainable Use License, which is fair-code rather than OSI open source.

What does this Railway template deploy?

Five services: n8n (editor and API, public domain, port 5678), n8n-worker (executes workflows), n8n-runner (isolated Code-node runtime from ghcr.io/n8n-io/runners:2.32.6-distroless), plus managed Postgres and Redis. Secrets are generated at deploy time and the service references are pre-wired.

Why does self-hosted n8n need Postgres and Redis?

Postgres holds workflows, encrypted credentials, execution history, community-node records, and binary data — it is the only durable store in the stack, which is why no n8n service needs its own volume. Redis is the job broker that queue mode requires; without it the main process would execute workflows itself and one heavy run could block the editor.

How do I scale n8n workers on Railway?

Increase throughput by raising replicas on n8n-worker. Note that n8n requires one task-runner sidecar per worker, so give each worker its own runner service pointed at that worker's broker on port 5679 rather than sharing a single runner.

Can I run Python code in self-hosted n8n?

Yes. N8N_NATIVE_PYTHON_RUNNER=true enables the native Python runner in the launcher, so a Code node set to Python executes in the runner container alongside the JavaScript runtime.

Do community nodes survive a redeploy without a volume?

Yes. Nodes installed through the UI are recorded in Postgres, and N8N_REINSTALL_MISSING_PACKAGES=true reinstalls them on boot, so they persist across container replacements.

Is n8n free for commercial use?

Yes for your own internal business automations. The Sustainable Use License only restricts redistributing or reselling n8n itself as a hosted product to third parties; running it to automate your own operations needs no licence purchase.