Choosing a hosting option
SpliceRun runs your automation apps three different ways. They are not tiers of the same thing - they suit genuinely different situations, and picking the right one at the start saves you moving later.
| Shared | Dedicated | Bring your own server | |
|---|---|---|---|
| Who owns the server | SpliceRun | SpliceRun, bought for you | You |
| Who it suits | Trying things out, small workloads | Production workloads | Teams with existing infrastructure |
| Resources | A fixed slice of a shared machine | A whole machine, yours alone | Whatever your server has |
| Neighbours | Other instances on the same host, isolated from each other | None | None |
| Server billing | Included | Included | You pay your provider directly |
| Support | Documentation and best-effort email | Included, one business day | Included, one business day |
| Backups | Daily, included | Daily, included | Daily, included |
Shared
Your instance runs in its own isolated container, with its own database and its own storage volume, on a machine shared with other SpliceRun customers. You get a fixed allocation of CPU, memory and disk.
What you actually get
These are enforced limits, not guidelines. Your container is capped at exactly these figures, which is also what stops a busy neighbour taking capacity from you:
| Shared XS | Shared S | |
|---|---|---|
| CPU | 0.5 vCPU | 1 vCPU |
| Memory | 1 GB | 2 GB |
| Disk | 10 GB | 20 GB |
| Price | $5/mo | $9/mo |
Prices are in US dollars. Indian customers are billed in rupees instead - see How pricing works.
A vCPU is a full core's worth of scheduling time. 0.5 vCPU means your instance can use half a core continuously, or burst against whatever is idle - n8n spends most of its life waiting on webhooks and HTTP calls, so a light-to-moderate workflow set fits comfortably.
Choose this if you are evaluating, running a handful of workflows, or want the lowest possible starting cost.
Consider something else if you need a lot of memory, run heavy or long workflows, or have a compliance requirement about shared infrastructure. A workflow that processes large files, holds big datasets in memory, or runs long CPU-bound loops will feel the 0.5 vCPU cap - Shared S doubles it, and Dedicated removes the ceiling entirely.
If your instance is being throttled you will see it as slow executions rather than errors. Why is my n8n slow? walks through how to tell the difference between a resource limit and a workflow problem.
Dedicated
A whole server, provisioned for you and used by nobody else. You choose the provider and the size; we buy it, set it up, and manage it.
Choose this if the automation matters to your business, you have real workflow volume, or you want the predictable performance of hardware nobody else is using.
Because you pick the provider, you can weigh cost against network quality and data location yourself rather than accepting one blended option.
Bring your own server
You already have a server - at any provider, or in your own rack. You add our SSH key, point a domain at it, and we install and run the app on your machine.
Choose this if you have infrastructure already, a data-residency requirement, an existing provider relationship or credits, or a policy that the primary copy of your data must stay on hardware you control.
Worth knowing before you choose it:
- The server is yours. Its operating system, disk, firewall and your relationship with your provider stay your responsibility. See What we manage.
- You need a domain you control, pointed at the server, before setup runs.
- We hold an SSH key on your machine so we can manage the app. You can revoke it at any time, and we remove it automatically when you cancel. See Data and access.
Full requirements: Bring your own server.
Moving between options
You are not locked in. Because every option runs the same application and the same backup format, moving is a restore into the new place rather than a rebuild - your workflows and credentials come with you.
Talk to us before you move so we can sequence it with minimal downtime.
Which is cheapest?
It depends on more than the SpliceRun price:
- Shared has the lowest total cost - one price, nothing else to pay.
- Dedicated costs more because it includes a machine nobody else uses.
- Bring your own server has the lowest SpliceRun fee, because we are not supplying hardware - but you pay your own provider separately. Whether that works out cheaper overall depends entirely on what you are already paying them.
Current prices are on the pricing page.