n8n on InstaPods vs Railway
Railway's Hobby plan reads as $5/mo n8n hosting. It is $5 of usage credit against a meter that charges $10 per GB of RAM per month, and n8n runs 24/7 by design. InstaPods is $7 flat, no meter.
n8n vs Railway: The Verdict
Pick Railway if you love the deploy experience, want templates and a managed Postgres one click away, and your services genuinely scale down. Pick InstaPods if n8n is always on, because a service that must stay up to catch webhooks is exactly the workload usage billing prices worst.
Last updated: 2026-09-02Quick Comparison
| Feature | Railway | n8n |
|---|---|---|
| Headline Price | $5/mo Hobby, includes $5 of usage | $7/mo flat, no meter |
| Memory Billing | $10 per GB per month | Included - 2 GB on Build |
| CPU Billing | $0.00000772 per vCPU per second (about $20/vCPU/mo) | Included - 2 vCPU on Build |
| Egress | $0.05 per GB | Included |
| Realistic n8n Cost | $20-25/mo at the 2 GB n8n realistically needs | $7/mo |
| Free Tier | $1/month of usage on the Free plan | No free tier - card required |
| Bill Predictability | Moves with load every month | Same number every month |
| PostgreSQL | One click, billed on the same meter | One click on the same pod, included |
| SSH Access | No - CLI shell into the build, not the box | Full SSH on every pod |
| Deploy Experience | Excellent - templates, PR environments, great DX | One click from the app catalog |
Where Railway's Meter Bites for n8n
Trade-offs to weigh before committing.
The $5 Plan Is Not a $5 n8n Bill
Hobby is $5/mo and includes $5 of usage credit. Railway bills memory at $10 per GB per month, so an n8n service sized to the 2 GB it realistically needs burns $20 of memory alone before a single vCPU-second. The credit covers about a week.
Always-On Is the Worst Case for Usage Billing
n8n has to be up to catch webhooks and fire cron triggers, so it bills 730 hours a month, every month. Usage pricing rewards workloads that idle at zero. This one never does, and you get none of the discount a flat plan gives for that commitment.
The Bill Moves With Your Workflows
A busy month of executions is a bigger invoice. That is fine for a product with revenue attached to the load, and uncomfortable for an internal automation box where you just wanted a predictable line item.
No Real Shell
Railway gives you logs and a CLI, not a server you can SSH into. When an n8n upgrade goes sideways at 2am, poking at the box is not an option.
Why Flat Pricing Suits an Always-On Tool
What you get when you pick this stack.
Flat $7, However Busy the Month Was
2 vCPU, 2 GB RAM and 25 GB of disk for $7/mo, whether n8n ran 10 executions or 10,000. No GB-hours, no egress line, no bill you have to model in advance.
Sized Right at the Entry Price
Build gives n8n the 2 GB its own docs ask for as the starting plan, not as an upgrade. On Railway that same 2 GB is $20/mo of metered memory.
Postgres Without a Second Meter
Add managed PostgreSQL in one click on the same pod. It is included in the $7, not another metered service with its own memory and CPU charges.
A Real Server With SSH
Every pod is a Linux server you can SSH into. Tail n8n logs, run the n8n CLI, take a database dump. Plus daily backups with 5 copies kept, and one-click upgrades that snapshot first and roll back on failure.
Real Cost Comparison
12 months of an always-on n8n sized to the 2 GB n8n realistically needs.
- Hobby plan base: $5/mo (includes $5 of usage)
- Memory, 2 GB always on: $20/mo at $10 per GB
- CPU: metered on top at $0.00000772 per vCPU per second
- Volume for the n8n data directory: billed per GB
- Egress: $0.05 per GB
- InstaPods Build plan: $7/mo (2 vCPU / 2 GB / 25 GB)
- Managed PostgreSQL on the same pod: included
- Bandwidth: included
- Daily backups, 5 copies kept: included
- $10 credit when you add a card (one-time)
Frequently Asked Questions
How much does n8n really cost on Railway?
The Hobby plan is $5/mo and includes $5 of usage, but usage is what you actually pay. Railway charges $10 per GB of memory per month and $0.00000772 per vCPU per second, which works out near $20 per vCPU per month. An n8n service sized at 2 GB and always on lands around $20-25/mo before egress and volume. Sized down to 1 GB you are nearer $10-15, at the cost of running below the memory n8n realistically needs.
Does Railway's free plan run n8n?
The Free plan includes $1 of usage per month and the trial gives a one-time $5 credit for 30 days. At $10 per GB-month of memory, $1 covers roughly three days of a 1 GB n8n service. It is enough to try n8n, not to run it.
Why is usage billing bad for n8n specifically?
Usage pricing pays off when a service idles at or near zero. n8n cannot: it has to be running to receive webhooks and fire scheduled workflows, so it bills all 730 hours in a month regardless of how many executions it handled. You get the volatility of metered pricing without the savings it is supposed to buy.
Is Railway better than InstaPods for anything?
Yes. Railway's deploy experience, templates and PR environments are genuinely excellent, and if you are shipping an app with a build pipeline rather than running a packaged tool, it is a nicer place to work. For a single always-on n8n instance where you want one predictable number, the flat $7 pod wins on both price and simplicity.
Can I move my n8n workflows from Railway to InstaPods?
Yes. Export the workflows as JSON from the n8n UI or CLI, deploy n8n on InstaPods in one click, and import them. Credentials are encrypted by n8n and need re-entering, but workflow structure transfers cleanly.
Deploy n8n in 60 Seconds
One number, every month, however busy your workflows get. $7/mo flat with Postgres and backups included.
Deploy n8n