Side by side
| VPS | PaaS | Serverless | Uno computer | |
|---|---|---|---|---|
| You rent | A virtual server | A place that runs your app | A function, per call | A microVM with root |
| You manage | The system, updates, security, the app | The app and its settings | The function | The app; an agent can do the system work |
| Billed by | The month or the hour | Resources, by the second or month | Requests and run time | A flat monthly plan |
| Longest run | No limit | No limit while the instance runs | AWS Lambda: 15 minutes per call | there is no session time limit; it sleeps when idle and wakes with its memory |
| State between requests | Disk and memory | Disk on a volume; memory restarts with the app | None: built for short tasks that keep no state between calls | Disk and memory |
| When nobody visits | Runs and bills | Runs; some free tiers spin down after 15 minutes | Nothing runs, nothing is billed | Asleep after 10 minutes, memory kept |
| Cost, 1 GB class | $6 (DigitalOcean) | $6.70 (Fly) to about $30 (Railway) | Free for light use; $43.20 if it ran all month | $5 |
A month of one 1 GB service that never stops
In $ per month. Serverless bills run time, so the last bar is the wrong use of it: a function that is idle most of the month costs almost nothing.
Show as a table
| Option | $ per month | Shape |
|---|---|---|
| Uno Small | 5 | 1 vCPU · 1 GB · sleeps when idle, keeps memory |
| VPS: DigitalOcean | 6 | 1 vCPU · 1 GB |
| PaaS: Fly shared | 6.70 | shared-cpu-1x · 1 GB, cheapest regions |
| PaaS: Railway | 30 | 1 vCPU · 1 GB used all month, billed by the second |
| Serverless: AWS Lambda | 43.20 | 1 GB running all month |
What each one looks like
With a VPS you log in and run commands. With a PaaS you push code and the platform builds and starts it. With serverless you write one function, and the provider calls it. Lambda allows 15 minutes per call, and 1,769 MB of memory gives one vCPU.
# VPS: you run the server $ ssh [email protected] $ apt install nginx $ systemctl enable --now myapp # PaaS (Fly.io): the platform runs it $ fly deploy # Serverless (AWS Lambda, Python): one function def lambda_handler(event, context): return {"statusCode": 200, "body": "ok"}
Serverless is cheap when quiet. Lambda gives 1 million requests and 400,000 GB-seconds a month free, then charges $0.20 per million requests. A 512 MB function that runs 200 ms a million times uses 100,000 GB-seconds, inside the free tier. A PaaS bills each piece: on Fly, a Machine, a volume and an IPv4 address are separate lines, and managed Postgres starts at $38 a month.
Who manages what
When nobody is there
Most apps spend most hours idle. How each model treats those hours decides the bill and the speed of the next visitor. A VPS keeps running. A serverless function does not exist between calls, so each new copy starts cold. An Uno computer sleeps after 10 minutes, and memory and running programs are saved while it sleeps.
Where Uno sits
An Uno computer is a VPS-class machine: a Firecracker microVM with root, one per computer. It has there is no session time limit; it sleeps when idle and wakes with its memory. It borrows the good parts of the other two. Like a PaaS, it can be set up for you, by an agent that installs and starts your app while you watch. Like serverless, it stops costing attention when nobody visits, but it keeps its state, so the next request finds everything where it was.
What an Uno computer adds to a plain VPS
Same root, same Linux. Here is what comes on top.
It sleeps when idle and keeps its memory
With nothing to do for 10 minutes it falls asleep with every program still in memory. A visitor, a bot message or you wake it. Every hour asleep earns you a boost.
An AI agent sets it up for you
Tell Uno what you need in a chat. It installs, configures and starts it on your computer, and you watch every command it runs. Or bring Claude Code or Codex with your own subscription.
Fast where it counts
NVMe disks and DDR5 with error correction (ECC). On 16 vCPU an Uno computer scores like a MacBook Air M4 in Geekbench 6: a second MacBook in the cloud for $70 a month.
A clean address of your own
Traffic can leave through a static address of a regular internet provider in one of 8 countries, used only by your account. Fewer captchas for your agent's browser.
Nothing happens unseen
Every time Uno itself touches your computer, it goes to your access log: who, when and why.
Where Uno is worse
- Serverless scales out to thousands of copies on its own. One Uno computer is one machine, up to 32 vCPU and 48 GB.
- A serverless bill falls to near zero when nobody calls. An Uno plan is flat, from $5 a month.
- A PaaS ships deploy pipelines, rollbacks and scaling as a product. On Uno you get a computer, and your agent builds those parts.
- A plain VPS sells memory for $1.25–1.62 per GB a month. Uno charges $4.2–5.
- Our vCPU are shared, and paid computers run in the Netherlands today.
An Uno computer, from $5 a month
Pick the size. It sleeps when idle and wakes on the first request. Move up any time.
Small
A Telegram bot, a VPN, a small backend
$5/ month
- 1 vCPU · 1 GB
- 15 GB disk
- 25 GB cloud storage
- 5 boosts a month
Plus
Agents that work with your laptop closed
$20/ month
- 4 vCPU · 4 GB
- 40 GB disk
- 100 GB cloud storage
- Uno Work in the cloud
Pro
A product in production, a second MacBook
$70/ month
- 16 vCPU · 16 GB
- 120 GB disk
- 300 GB cloud storage
- Speed of a MacBook Air M4
Max
A team and many agents at once
$200/ month
- 32 vCPU · 48 GB
- 300 GB disk
- 1,024 GB cloud storage
- 5 public IPv4
Add Uno AI to any plan: 40 h of AI work a month on Small, 100 h on Plus, unlimited on Max. Today: GLM-5.3 Flash and MiMo v2.6 Pro. Or sign in with your own Claude or ChatGPT.