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

# AWS CLI and boto3

The [AWS CLI](https://aws.amazon.com/cli/) is the command-line client for S3, and [boto3](https://boto3.amazonaws.com/v1/documentation/api/latest/index.html) is the Python library underneath it. Both work against a Tapedrive bucket once a profile points them at the gateway.

## Before you start

You need a bucket and a credential. A bucket is a tape that has delegated writes to the gateway, and a credential is an access key pair the gateway operator issues you. [Set up a bucket](/tools/s3-gateway#set-up-a-bucket) walks through both.

## Configure

Add a profile for the gateway:

```ini ~/.aws/config theme={null}
[profile tapedrive]
region = solana
endpoint_url = https://s3.example.com
```

```ini ~/.aws/credentials theme={null}
[tapedrive]
aws_access_key_id = TAPE4Q7ZK2M9XH3PD8AW
aws_secret_access_key = mZ9vQ3nL8kR2tW6yB4cX7dF1hJ5pS0aG3uE8iO2q
```

The gateway ignores the region, so any label works. These pages use `solana`.

## Everyday commands

```bash theme={null}
# list your buckets
aws s3 ls --profile tapedrive

# list objects under a prefix
aws s3 ls s3://my-bucket/photos/ --profile tapedrive

# upload a file
aws s3 cp ./photo.jpg s3://my-bucket/photos/photo.jpg --profile tapedrive

# upload a directory
aws s3 sync ./photos s3://my-bucket/photos/ --profile tapedrive

# download a file
aws s3 cp s3://my-bucket/photos/photo.jpg ./photo.jpg --profile tapedrive

# delete an object
aws s3 rm s3://my-bucket/photos/photo.jpg --profile tapedrive

# a link anyone can open for an hour
aws s3 presign s3://my-bucket/photos/photo.jpg --expires-in 3600 --profile tapedrive
```

Large files switch to multipart upload on their own.

## boto3

The same three values build a client:

```python theme={null}
import boto3

s3 = boto3.client(
    "s3",
    endpoint_url="https://s3.example.com",
    aws_access_key_id="TAPE4Q7ZK2M9XH3PD8AW",
    aws_secret_access_key="mZ9vQ3nL8kR2tW6yB4cX7dF1hJ5pS0aG3uE8iO2q",
    region_name="solana",
)

# upload and download
s3.upload_file("photo.jpg", "my-bucket", "photos/photo.jpg")
s3.download_file("my-bucket", "photos/photo.jpg", "photo.jpg")

# list a prefix, one page at a time
for page in s3.get_paginator("list_objects_v2").paginate(Bucket="my-bucket", Prefix="photos/"):
    for item in page.get("Contents", []):
        print(item["Key"], item["Size"])

# delete
s3.delete_object(Bucket="my-bucket", Key="photos/photo.jpg")
```

A 24 MiB upload goes as a three-part multipart upload in about 2 s and downloads in under a second.

## Good to know

* You cannot create a bucket through S3. Create the tape with [`tape create`](/tools/cli/commands/tape-management#create) and delegate it to the gateway.
* ETags are not MD5 hashes. Tools that compare an ETag with a local MD5 report a mismatch that is not corruption ([compatibility](/tools/s3-gateway/compatibility)).
* Every object write is a transaction on the chain. A bucket takes one write per block, so many small uploads to one bucket are paced by the block time, not by bandwidth.

## Next

* [Uploads](/tools/s3-gateway/uploads) for multipart uploads and overwrites.
* [Integrations](/tools/s3-gateway/integrations) for the errors every S3 client shares.