Streamline Your AI Workflows with RunPod’s New S3-Compatible API

Managing files in your Runpod network volumes just got a whole lot easier.

Today we’re excited to announce the launch of our new S3-compatible API, a powerful way to access and manage files on your Runpod volumes without launching a Pod. If you’ve ever spun up a GPU just to check a file or move some data around, this feature is for you.

Our mission is to break down barriers to great ideas. Now you can manage your data with ease, create faster, and build better AI.

What is the S3-compatible API?

Runpod now offers an API that uses the same syntax as the S3 protocol, giving you direct access to your network volumes (your persistent storage on Runpod) using the tools you already know, like the AWS CLI and the Boto3 Python SDK. It supports all the essentials: listing, uploading, downloading, deleting, and syncing files.

This new API gives you:

• Direct file access without launching a Pod.
• Lower operational overhead for managing storage.
• Standard S3 tooling support.

It’s the same underlying network storage your Pods already mount, just now accessible with S3.

When to use S3-compatible API vs. editing files with Pods

If you're wondering how this differs from mounting volumes directly inside Pods, here's the distinction:

• Use S3-compatible access when you want fast, compute-free interaction with your files — for example, to upload training data, archive model outputs, or sync datasets across systems without spinning up a GPU.
• Use Pod-based access when your code needs low-latency, in-place read/write access during runtime — for example, while training a model or doing inference. Remember, a drive connected directly to the machine is always going to be the fastest, highest throughput method of storage, so for tasks that are constantly interacting with the drive they will be preferred.

Both access methods point to the same storage, so you can fluidly move between them depending on your workload. This allows you to always rely on the right tool for the task at hand: whether you need accessibility or speed, you now have access to both.

Why it matters

Before today, managing files often meant launching a Pod, even for simple tasks like checking a directory or copying a dataset. That added friction, not to mention cost.

Now, with the S3-compatible API, you can:

• Inspect and manage storage instantly from your terminal.
• Use the AWS CLI or Boto3 to automate file flows across pipelines.
• Streamline operations without needing an active Pod.

It’s the same secure storage system, now accessible in a simpler, more flexible way. No more paying $0.50/hour for a GPU and spending five to ten minutes just to upload a 10GB dataset.

How it works

Getting started is straightforward:

1. Generate an S3 API key from the Runpod console.
2. Configure your AWS CLI or script environment with that key.
3. Use aws s3 commands to access your volume, passing in the Runpod datacenter endpoint URL.

For this initial release, the S3-compatible API is available for network volumes in these four datacenters:

• Datacenter: EUR-IS-1, Endpoint URL: https://s3api-eur-is-1.runpod.io/
• Datacenter: EU-RO-1 , Endpoint URL: https://s3api-eu-ro-1.runpod.io/
• Datacenter: EU-CZ-1 , Endpoint URL: https://s3api-eu-cz-1.runpod.io/
• Datacenter: US-KS-2 , Endpoint URL: https://s3api-us-ks-2.runpod.io/

You’ll find the datacenter ID for your existing network volumes in the Storage section of the Runpod Console.

Unlike traditional object storage, Runpod’s S3-compatible API maps files directly to your volume’s file system. What you see is what’s actually there.

For a complete walkthrough of setup and CLI/Boto3 usage examples, check out the Runpod documentation.

Get started today

Whether you're syncing datasets, archiving checkpoints, or just want a quicker way to inspect your storage, the S3-compatible API is ready to go.

Head to your Runpod console to generate an S3 API key and start exploring, or dive into the docs for usage examples.