Customer A
Scoped access- app
- jobs
2 / 4 VMs
- CPU
- 2 / 4 vCPU
- RAM
- 2 / 8 GiB
- Disk
- 20 / 40 GiB
Machines for humans.
And their agents.
Full computers in the cloud for development, apps, and AI agents.
Start one from the dashboard or create a fleet through the API.
Start a VM
Illustrated Linux VM workflow: allocate compute and memory, boot Linux, connect the runtime, and execute a command.
Docker
Python
Node.js
PostgreSQL
Go
Redis
Bun
Rust
Kubernetes
git
TypeScript
NGINX
Ubuntu
Deno
.NET
Ruby
A machine for
every agent.
Give each agent a computer for its code, tools, and files. Run VMs in parallel, connect them when needed, and manage the fleet through one API.
An illustrated fleet of independent Linux VMs. Add machines to grow the fleet, or select a machine to highlight its connection. These controls change the example only.
client.vms4 machines running
Full control. Built in.
Everything you need to work in a machine, or build with a fleet.
Runtime API
Commands, files, and processes. In code.
SSH access
Your machine. Your familiar tools.
Browser desktops
Open a computer in your browser.
File tools
Browse, edit, and move your files.
Root access
Install your stack. Make it your own.
Custom images
Your OS. Your tools. Your starting point.
Private networking
Connect only the machines you choose.
HTTPS previews
An HTTPS URL for your running app.
Custom domains
Your app, at your own address.
Managed workloads
Keep services running. Follow their logs.
Team access
Shared machines. Individual permissions.
Scoped API keys
Give automation the access it needs.
Persistent files
Stop the machine. Keep your files.
Temporary VMs
A lifetime for every task.
Sleep & wake
Sleep when idle. Wake on a request.
Disk archives
Save a backup. Come back to it later.
Resource pools
One pool. A whole fleet of possibilities.
Audit history
See who changed what, and when.
From a quick task
to a connected fleet.
Temporary for a job. Persistent for your apps. Connect your VMs, share capacity, and manage their lifecycles from one dashboard and API.
| What you want to do | Traditional VPS / serverA typical self-managed setup |  Lifecycles, networking, and fleets built in |
|---|---|---|
Short-lived jobs | Write cleanup scripts—or do it yourself Track expiry times, find finished jobs, and stop or delete the servers they leave behind. | Set an expiry. Oblien handles it. Choose stop or removal at expiry, or automatic removal on exit. Extend the lifetime when needed. |
Long-running apps | Persistent servers Install your stack and configure service startup and recovery yourself. | A computer for the long run Choose persistent mode, keep your files, and set a VM restart policy. |
Changing lifetimes | Maintain your own expiry logic Add, change, or remove cleanup jobs whenever a server’s role changes. | Change the mode. Keep the VM. Switch between temporary and persistent. Keep the same computer and files without rebuilding. |
Idle workloads | Runs until you stop it Stays running when your app is idle. | Always on, or sleep when idle Opt into sleep on compatible Linux VMs. Save the session and wake on a web request. |
VM-to-VM connections | Manual setup—or scripts you maintain Configure the network, allowed sources, and ports. Keep the rules in sync as servers change. | Choose which VMs can connect Select the VMs and ports. Oblien resolves VM addresses and applies their private access rules. |
Fleet management | Provision servers. Build the fleet logic. Choose instance sizes, configure access, and maintain scripts for group limits, stop actions, and cleanup. | Fleet controls are already built in Create VMs from a shared pool. Group them, set resource caps, and stop a group with one call. |
Flexible capacity | Capacity tied to individual servers Pick and provision each instance size. Resize or replace instances as your needs change. | Buy a pool. Split it across VMs. Allocate CPU, RAM, and storage to one VM or many. Replace VMs within the same pool. |
Automation | Maintain scripts across APIs and SSH Write the code for server creation, remote commands, file transfers, and cleanup. | One SDK from creation to cleanup Create VMs, run commands, manage files, and connect services through the same interface. |
Run and publish apps | Install and maintain supporting tools Set up process supervision, logs, a proxy, and TLS—or add separate managed services. | Choose a command. Expose a port. Workload controls, streamed logs, and HTTPS are included. Add your domain with DNS setup. |
Typical self-managed workflows; providers also offer APIs, consoles, private networks, and snapshots. Sleep/wake requires a compatible Linux image, available capacity, and credit. Retained storage remains billable; account, image, and per-VM limits apply. Custom domains require DNS setup.
Your product. Their computers.
Give every customer their own VMs, access, and resource limits. One namespace per customer. One API for your whole product.
Your applicationOne API for every customer
Customer B
Scoped access- app
- jobs
- data
3 / 6 VMs
- CPU
- 4 / 8 vCPU
- RAM
- 8 / 16 GiB
- Disk
- 40 / 80 GiB
Your next customer
Their own VMs, access, and limits.
Customer C
Scoped access- app
- jobs
2 / 4 VMs
- CPU
- 2 / 4 vCPU
- RAM
- 2 / 8 GiB
- Disk
- 20 / 40 GiB
2 customer namespacesAllocated / customer limit
Price the fleet you need.
Choose how many VMs can run together, then set the resources for each. That’s your shared pool. Choose how to pay for it.
Build your fleet
Your shared pool
Capacity for 4 VMs of this size at once.
- CPU
- 2 vCPU
- RAM
- 4 GiB
- Disk
- 40 GiB
Pay as you go
Set your maximum capacity. Pay for what you use.
$0.044/ compute hour
Whole fleet at 20% CPU activity, with its RAM reserved. Adjust your usage below.
- Retained disks for this fleet
- $2 / month
- Monthly resource ceiling
- $22.50
No upfront pool fee. CPU, RAM, and disk charges stop at 125% of the monthly pool price, even at full usage.
Start with PAYGManaged internet transfer: $0.10/GiB, billed separately.
Monthly pool
A fixed price for resources you use every day.
$18/ month
Prepay for the whole pool. Run your fleet throughout the paid month.
- CPU, RAM, and disk
- Covered for the month
- Managed internet transfer
- 100 GiB included
Split, replace, or resize VMs within the same pool. The price belongs to the resources, not individual machines.
Choose monthlyOptional transfer overage: $0.10/GiB. Off by default.
Pay as you go. Never past your ceiling.
CPU, RAM, and disk charges stop at your pool’s monthly limit—even at full usage.
Estimated pay as you go
$12.56/ month
Save $5.44vs the monthly pool
Monthly resource ceiling$22.50
0 h/day12 h/day24 h/day
- Active CPU
- $2.88
- Reserved RAM
- $7.68
- Retained disk
- $2
30-day estimate. RAM billed while reserved; disks kept all month. Transfer and other services extra.
USD resource estimates. Account plans, image requirements, and available host capacity apply. Confirm your pool and final price in the dashboard.
How the pool and billing work
A pool, many possible fleets
The VM count is a starting layout. You can mix smaller and larger VMs within your CPU, RAM, and disk pool. Each VM must meet its image’s minimums and platform limits; current account and VM-count policies still apply.
Small Linux VMs can start at 0.1 vCPU, 128 MiB RAM, and 64 MiB writable disk. Images can need more. Fractional CPU is a share of CPU time. A pool starts at 256 MiB RAM and 1 GiB storage, with storage rounded up to whole GiB.
What PAYG measures
$0.03 per active vCPU-hour, $0.008 per reserved GiB-hour of RAM, and $0.05 per retained GiB-month of storage. The hourly figure follows the CPU activity selected above. Activity reduces the CPU charge, not reserved RAM or retained storage.
A confirmed stop releases running CPU and RAM. Paused RAM and retained disks remain billable. The monthly ceiling covers resource usage at the chosen pool size; managed transfer and other services are separate. PAYG requires a funded account balance.
What monthly covers
The fixed price covers the purchased CPU, RAM, and disk for the paid month, plus the shown shared transfer allowance. Stopping a VM does not reduce that prepaid fee. Retained disks remain billable after coverage ends. Both billing modes use account credits; account plans set separate resource limits.
Public routes and your own proxy do not incur the managed-proxy charge. Backups and applicable taxes are separate. Larger fleets may need a custom capacity review.
Published rates Billing details Rates checked 2026-11-12
A few things
to know.
Talk to us Can I use Oblien as my own cloud computer?
Yes. Create a VM from the dashboard, connect over SSH, or open a desktop in your browser on a supported image. Install your tools, develop remotely, and keep files on a persistent machine. You can also host apps, databases, and background jobs. Explore remote desktops.
Can my product create machines for its users?
Yes. Use the API to create a VM for each user, agent, or task. Run them in parallel, use temporary machines for short tasks, or keep persistent machines between sessions. Your backend can run commands and manage files through the Runtime API. Explore the use cases.
Which operating systems can I run?
The catalog includes Linux, macOS, Windows and Android environments. Resource requirements, desktop access and lifecycle features depend on the image. Explore images and presets.
What is a microVM?
A microVM is a lightweight virtual machine. Oblien’s Linux machines use Firecracker, with a separate kernel and memory space for each VM. Each agent or workload gets its own environment for code, tools, and files, managed through the same dashboard and API. See how isolation works.
What happens when I stop a VM?
A normal stop keeps its writable disk, so your files are there when you start it again. Live process memory is not retained by a stop. Retained storage remains billable; running-capacity rules depend on your pool’s allocation basis. See how capacity works.
Can my machines connect to each other?
Yes, through private links and explicitly opened service ports. Links allow connections in one direction; inbound access starts closed. Runtime API requests also require authentication. Read about VM networking.
How does sleep and wake work?
On a compatible Linux VM, idle sleep saves memory to disk and releases the running VM process. A request to its exposed web app restores it. Memory restore support varies by image; nested KVM environments such as the current Mac image do not support it. Explore sleep and wake.
Do I pay for each machine separately?
You purchase a pool of CPU, memory and storage that can be split across your VMs, within your account and image limits. Choose prepaid monthly capacity or usage-based billing with a cap. Network transfer and retained storage follow their own billing rules. Read the billing details.
Your next idea.
Give it a machine.
Start with one machine. Build a fleet when you need it.