This website uses cookies to anonymously analyze website traffic using Google Analytics.
Company

Together Code Sandbox: the most robust infrastructure for building AI coding products at scale

May 20, 2025

By 

Together AI

The world runs on code, and most of that code is now being generated by LLMs. Together Code Sandbox provides a fast, reliable way to build and scale the next generation of AI-powered coding tools.

Powered by a robust infrastructure, Together Code Sandbox provides programmatic access to the leading virtual sandboxing product, allowing developers to spin up sandboxes in milliseconds and use a wide range of CPU compute (up to 64 vCPUs and 128 GB RAM per sandbox) to build full-fledged coding environments or set up powerful agentic workflows.

With this launch, developers can now access cutting-edge technologies such as memory snapshotting and a suite of mature developer tools and APIs, empowering them to rapidly create innovative AI products.

Building code tooling in the AI era

The past year has seen explosive growth in AI-driven coding solutions, from intelligent coding assistants to fully autonomous agents. But as these tools become more sophisticated, they require the underlying infrastructure to be able to run AI-generated code securely, efficiently, and at scale.

Together Code Sandbox was specifically built for these requirements. With instant VM startup, effortless scalability to thousands of concurrent sandboxes, and proven reliability, you can execute complex AI-generated workflows confidently. Spend less time managing infrastructure and more time creating innovative AI-powered experiences.

Fast, secure sandboxes for scale

One of our top priorities while building Together Code Sandbox was to deliver exceptionally fast sandbox spin-up times.

By leveraging CodeSandbox’s proprietary implementation of Firecracker microVM technology, Together Code Sandbox starts VMs from a snapshot in 500 ms and creates VMs from scratch (cold start) in 2.7 seconds (P95).

Understanding the multitude of use cases that Together Code Sandbox can address, we have introduced several customization options and a suite of dev tools to make it a highly flexible solution:

  • Scalable snapshotting solution: Tried-and-true memory snapshotting technology that hibernates sandboxes and resumes them from a snapshot.
  • Interactive terminals & port previews: Real-time shell interactions and immediate access to running applications.
  • Configurable tasks: Customize background processes and automated workflows.
  • Flexible sessions: Supports multiple users simultaneously for real-time collaboration.
  • VM scaling: Hot-swappable VM sizes ranging from 2 to 64 vCPUs and 1 to 128 GB RAM.
  • Customizable environment: Go beyond Docker images by leveraging our Dev Container environment to specify an environment for your use case.
  • Persistent storage: A git-versioned filesystem ensures your data remains accessible and secure.

To empower developers working on projects of any scale, we are launching Together Code Sandbox with transparent, competitive pay-as-you-go pricing:

  • $0.0446 per vCPU per hour, plus
  • $0.0149 per GiB RAM per hour.

You can find more information about available VM sizes in our docs.

Try our Together Code Sandbox playground for an interactive walkthrough of the different capabilities (or catch a sneak peek below).

Tried and tested at scale

Together Code Sandbox was refined through extensive collaboration and feedback from leading AI developers behind some of today’s most popular text-to-app builders and agentic frameworks.

While developing this technology, which was originally serving 4 million monthly CodeSandbox users, we quickly learned that our customers have the same needs as the ones that inspired our design of the CodeSandbox infrastructure. Many of these teams previously struggled with limited customizability, scalability bottlenecks, and inconsistent reliability from existing solutions. Traditional infrastructure struggled under the weight of hundreds or thousands of concurrent sandboxes.

Thanks to infrastructure optimzed over the last 3 years, Together Code Sandbox offers unmatched reliability and scalability. Our sophisticated memory snapshotting technology enables resuming VMs from hibernation in 511 milliseconds (P95), and the fully customizable environment (powered by Dev Containers) allows developers to easily configure any type of setup, including servers, databases, and custom Docker configurations.

Powering AI pioneers

Some AI pioneers, including Blackbox AI, Superblocks, and HeroUI, have already been testing this flexible infrastructure to accelerate their go-to-market and drastically improve code environment performance.

Thanks to Together Code Sandbox, HeroUI cut the development time of HeroUI Chat from 5 months to 2 weeks, reducing their VM startup times from 2 minutes to under 2 seconds.

“With Together Code Sandbox, we had an MVP running in a couple of hours, so it finally allowed us to test the result of HeroUI Chat, instead of waiting for months to see the result in a preview environment. Now, this is the core of HeroUI Chat; it is the platform that allows us to run the projects. Without it, we don't have a way to show you the result of the AI.” — Junior Garcia, Founder and CEO of HeroUI

Get started

Together Code Sandbox is production-ready and you can start building your apps with it today by following our Docs

If you’re interested in learning more about how you can leverage Together Code Sandbox at scale, contact our Sales team.

  • Lower
    Cost
    20%
  • faster
    training
    4x
  • network
    compression
    117x

Q: Should I use the RedPajama-V2 Dataset out of the box?

RedPajama-V2 is conceptualized as a pool of data that serves as a foundation for creating high quality datasets. The dataset is thus not intended to be used out of the box and, depending on the application, data should be filtered out using the quality signals that accompany the data. With this dataset, we take the view that the optimal filtering of data is dependent on the intended use. Our goal is to provide all the signals and tooling that enables this.

No items found.
Start
building
yours
here →