# InsForge Storage Now Supports the S3 Protocol

24 Apr 2026 • 4 minutes

Junwen Feng

Engineer

InsForge Storage now speaks the [AWS S3 protocol](https://docs.aws.amazon.com/AmazonS3/latest/API/API_Operations_Amazon_Simple_Storage_Service.html) natively.

Point any SigV4-signing client — the `aws` CLI, AWS SDKs, Terraform — at `/storage/v1/s3` on your project, hand it a project access key, and you're reading and writing the same buckets you already use through the REST API and the Dashboard.

## What It Unlocks

- **CI and build artifacts.** Push releases from GitHub Actions with `aws s3 cp` or `rclone sync` — no InsForge SDK wrapper needed.
- **Existing tooling, unchanged.** Terraform's `aws_s3_object`, backup scripts, log shippers, and DR pipelines keep working against a different endpoint.
- **Server-side uploads.** Workers, cron jobs, and batch scripts get storage access through credentials their runtimes already handle natively.

For browser uploads, signed download URLs, and bucket visibility, stay on the [TypeScript SDK](https://docs.insforge.dev/sdks/typescript/storage). S3 keys are long-lived and project-wide; the SDK is what you want in app code.

## Endpoint and Region

Both values live under **Storage → Settings → S3 Configuration** in the dashboard, or come back from `GET /api/storage/s3/config`.

| Field | Value |
| --- | --- |
| Endpoint | `https://{project-ref}.{region}.insforge.app/storage/v1/s3` |
| Region | `us-east-2` |

Clients must use path-style URLs (`forcePathStyle: true`). Virtual-hosted style is not supported.

## Access Keys

Generate credentials from **New access key** in the same dashboard panel, or via the admin API:

bashCopy

```bash

```

The secret is returned **exactly once**. It's encrypted at rest and we can't recover it later — capture it immediately, and revoke + recreate if you lose it.

## A Quick Example

AWS CLI — drop this into `~/.aws/config`:

iniCopy

```ini

```

Then:

bashCopy

```bash

```

The same endpoint and addressing-style flags work for , , and `rclone`. The [S3 gateway docs](https://docs.insforge.dev/storage/s3-gateway) have drop-in configs for each.

## Supported Operations

The gateway implements what real SDKs and tools actually call:

- **Bucket:** `ListBuckets`, `CreateBucket`, `DeleteBucket`, `HeadBucket`, `ListObjectsV2`
- **Object:** `PutObject`, `GetObject` (with `Range`), `HeadObject`, `DeleteObject`, `DeleteObjects`, `CopyObject`
- **Multipart:** full lifecycle — `CreateMultipartUpload`, `UploadPart`, `CompleteMultipartUpload`, `AbortMultipartUpload`, `ListParts`

Streaming uploads work by default. `aws s3 cp` with multi-GB files and `aws s3 sync` run without any client-side config changes.

## Get Started

The gateway is available on InsForge Cloud projects running **v2.0.9 or newer**. If your project is on an older version, upgrade it to the latest release first.

1. Open **Storage → Settings → S3 Configuration** in the dashboard
2. Copy the **Endpoint** and **Region**
3. Click **New access key** and save the secret
4. Point any S3 client at that endpoint with the key

- [S3 Gateway documentation](https://docs.insforge.dev/storage/s3-compatibility)
- [TypeScript storage SDK](https://docs.insforge.dev/sdks/typescript/storage)