How it works
You create a bucket, then upload objects into it. Each object has a key, a unique name inside the bucket such as photos/2026/cat.jpg. The slashes only look like folders. Underneath, the bucket is a flat list of keys, and "listing a folder" means "list keys that start with this prefix".
Everything is an HTTP request signed with an access key. That is why one protocol works from a laptop, a server and a phone, and why any tool that speaks S3 can use any S3-compatible storage. You change the endpoint address and the keys; the commands stay the same.
$ aws s3 cp backup.tar.gz s3://backups/db/ --endpoint-url $S3_ENDPOINT upload: ./backup.tar.gz to s3://backups/db/backup.tar.gz $ aws s3 ls s3://backups/db/ --endpoint-url $S3_ENDPOINT 2026-10-16 09:12:44 18204391 backup.tar.gz $ aws s3 presign s3://backups/db/backup.tar.gz --expires-in 3600 --endpoint-url $S3_ENDPOINT https://… a link that works for one hour
Objects are written and read whole. To change one byte in a large file you upload the file again. In return you get simple rules: the same key gives the same object. Amazon S3 itself also guarantees that a read after a successful write returns the new data.
Object storage, disk and database
| Computer disk | Object storage | Database | |
|---|---|---|---|
| You reach it by | A file path | Bucket and key, over HTTPS | A query |
| Edit a file in place | Yes | No, upload it again | Update a row |
| Best for | Code, packages, running programs | Backups, uploads, images, exports | Records you search and join |
| Tied to one machine | Yes | No | Usually |
| Share with a link | No | Yes, a temporary link | No |
A common design uses both: the database keeps a small row with the key, and the file's bytes live in a bucket. That keeps the database light.
The disk of a VPS is the first column. It is fast and simple, and it goes away with the server. On Uno the disk belongs to the computer, while cloud storage belongs to your account. How Uno storage works shows the difference in one picture, and this comparison puts the machine types side by side.
Do you need it?
Take object storage when
- Files must outlive a machine. Backups, exports, what an agent produced.
- Users upload things. Photos, PDFs, recordings: far too big for a database.
- Several machines read the same files. One bucket, many readers.
Skip it when
- A program edits the file in place. A running database or a compiler wants a disk.
- You need folders with permissions. A file server does that better.
- You have a few small files. A repository holds them.
Object storage in Uno
Every plan includes cloud storage. It belongs to your account, so it stays when you delete a computer. You reach it from Uno Work (Files, Cloud storage), the console, your agent over API and MCP, and presigned links. S3 access keys for your own tools are not available yet, so the AWS CLI or rclone cannot connect to it directly.
Small: 25 GB from $5
More room for your files
Storage grows with the plan. Delete a computer and the files stay.
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.