Storage
What it is
nodabase.storage is object storage for your application: upload a file, download it, check whether it exists, delete it. Nodabase uses S3-compatible object storage backed by MinIO underneath, and the client speaks the standard S3 protocol to it.
Files live under keys in one configured bucket — avatars/me.png, reports/2026.csv — so a whole application's objects are organized by key inside that bucket.
How it works
Node.js application
|
v
nodabase.storage
|
v
MinIO
(S3-compatible)Every call goes to the configured S3 endpoint with the configured credentials. Because the protocol is standard S3, the same bucket can be reached with any S3-compatible tool, not just this client.
Configuration
Five variables configure the module:
NODABASE_STORAGE_ENDPOINT=http://localhost:9000NODABASE_STORAGE_REGION=us-east-1NODABASE_STORAGE_ACCESS_KEY=your-access-keyNODABASE_STORAGE_SECRET_KEY=your-secret-keyNODABASE_STORAGE_BUCKET=my-app
NODABASE_STORAGE_ENDPOINT and NODABASE_STORAGE_BUCKET are required — without an endpoint, nodabase.storage is not created. The region defaults to us-east-1. When your application runs inside the Nodabase platform, all five bindings are provided for you.
Basic usage
import { nodabase } from "nodabase";await nodabase.storage.createBucket(); // ignore the error if it already existsconst result = await nodabase.storage.upload("avatars/me.png", buffer, {contentType: "image/png",});// { key: "avatars/me.png", bucket: "my-app" }
The body is a Buffer, string or readable stream — anything an S3 put accepts. upload resolves to the key and bucket it wrote to.
Common operations
The whole surface at a glance:
| Method | Returns | Notes |
|---|---|---|
createBucket() | true | creates the configured bucket; errors if it exists |
upload(key, body, { contentType }) | { key, bucket } | contentType is optional |
download(key) | readable stream | pipe it — not a Buffer |
exists(key) | true / false | |
delete(key) | true |
Downloading
const body = await nodabase.storage.download("avatars/me.png");body.pipe(res);
download resolves to a readable stream, not a buffer. Pipe it to a response or a file, or consume it with your stream tooling of choice.
Checking and deleting
await nodabase.storage.exists("avatars/me.png"); // true or falseawait nodabase.storage.delete("avatars/me.png"); // true
exists answers with a head request, so a false means no object under that key — there are no folders, only keys.
Important notes
createBucket()fails if the bucket already exists — that is normal; call it once during setup and ignore the error if it comes back.- The client exposes these five operations and no more. There is no presigned-URL API, no object listing and no copy in the client — S3-compatible tooling can talk to the same bucket directly when you need more.
downloadreturning a stream is the part people trip on: awaiting a Buffer does not happen automatically.- Without
NODABASE_STORAGE_ENDPOINT,nodabase.storageis absent from the client — the same pattern asdbandcache.
Next steps
- Cache — Redis-compatible caching through Dragonfly
- Queues — publishing and consuming through NATS
- Overview — how the client and the services fit together
- Quickstart — run the platform