> ## Documentation Index
> Fetch the complete documentation index at: https://docs.onebucket.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Quickstart

> Connect an S3 client or Claude and read and write your first object in minutes.

Get data flowing through OneBucket — as S3, as MCP, or both.

## Prerequisites

* A OneBucket account — [request one](https://app.onebucket.io/sign-up).
* From your [OneBucket console](https://app.onebucket.io): your **S3 endpoint** and an **access key / secret key** pair.

<Note>
  OneBucket is deployed per cluster, so your endpoints look like `https://s3.<your-cluster>.onebucket.io`. Copy the exact values from your console.
</Note>

## Use it as S3

OneBucket speaks the S3 API with AWS SigV4 auth, so any S3 client or SDK works — just change the endpoint.

<Tabs>
  <Tab title="AWS CLI">
    ```bash theme={null}
    # Credentials
    aws configure set aws_access_key_id <ACCESS_KEY>
    aws configure set aws_secret_access_key <SECRET_KEY>
    export OB=https://s3.<your-cluster>.onebucket.io   # copy from your console

    # Create a bucket, upload, list, and read back
    aws --endpoint-url $OB s3 mb s3://my-first-bucket
    echo "hello from onebucket" > hello.txt
    aws --endpoint-url $OB s3 cp hello.txt s3://my-first-bucket/hello.txt
    aws --endpoint-url $OB s3 ls s3://my-first-bucket/
    aws --endpoint-url $OB s3 cp s3://my-first-bucket/hello.txt -
    ```
  </Tab>

  <Tab title="boto3 (Python)">
    ```python theme={null}
    import boto3

    s3 = boto3.client(
        "s3",
        endpoint_url="https://s3.<your-cluster>.onebucket.io",  # copy from your console
        aws_access_key_id="<ACCESS_KEY>",
        aws_secret_access_key="<SECRET_KEY>",
        region_name="us-east-1",
    )

    s3.create_bucket(Bucket="my-first-bucket")
    s3.put_object(Bucket="my-first-bucket", Key="hello.txt", Body=b"hello from onebucket")
    print(s3.get_object(Bucket="my-first-bucket", Key="hello.txt")["Body"].read().decode())
    ```
  </Tab>
</Tabs>

Because it's standard S3, the same endpoint and keys work with SDKs in any language, `s3cmd`, rclone, and most S3-compatible tools.

## Connect Claude

Prefer to work in natural language? Add OneBucket as an MCP connector and let Claude list, read, and manage your objects for you.

<Card title="OneBucket Connector for Claude" icon="plug" href="/connectors/claude">
  Add your `mcp-core.<your-cluster>.onebucket.io` endpoint as a custom connector, sign in, and ask Claude to work with your buckets.
</Card>

## Manage keys and tokens (CLI)

The `onebucket` CLI handles authentication and tokens — your data operations go through the S3 client above.

```bash theme={null}
onebucket login      # OAuth browser sign-in; exports ONEBUCKET_TOKEN
onebucket keys       # list your access keys
onebucket token      # mint an MCP JWT for a client
onebucket sessions   # list or revoke active sessions
onebucket logout     # clear the stored login
```

## Next steps

* **Automate with AI** — [connect Claude](/connectors/claude) for natural-language access to your storage.
* **Span backends** — copy, migrate, and locate objects across clouds, on-prem, and edge from the same client.
* **Cut egress** — prefetch hot objects into the edge cache so repeat reads are served locally.