CubeSandbox

CubeSandbox is a sandbox service that runs AI agent code in isolated virtual machines for developers who need safe code execution.

What it is good for

  • Running untrusted code from AI agents safely
  • Starting many isolated code environments quickly
  • Controlling which websites and credentials an agent can reach

At a glance

Who it is for
  • Developers
Setup
Requires programming

Level 5 of 5. You need to write code to use it.

Cost
Free and open source
Type
Other
Tasks
Works with
E2B SDK, Terraform, Kubernetes

Momentum

Rank this week
Not ranked yet
Best rank
None yet
GitHub stars
13,131

GitHub numbers as of Nov 2, 2026. Updated weekly.

12.6k 12.9k 13.2k Sep 24, 2026 Nov 2, 2026
GitHub stars for CubeSandbox: 12,620 on Sep 24, 2026, 13,131 on Nov 2, 2026. Source: GitHub.

Official links

Setup steps are on the official pages. We do not repeat them here, since they change.

Questions

Is CubeSandbox free?
Yes. It is free and open source.
Do I need to know how to code to use CubeSandbox?
Yes. You need to write code to use it.
What does CubeSandbox work with?
According to its own documentation, it works with E2B SDK, Terraform, and Kubernetes.
What is CubeSandbox good for?
Running untrusted code from AI agents safely, starting many isolated code environments quickly, and controlling which websites and credentials an agent can reach.

Similar agents

  • CLI-Anything

    CLI-Anything is a toolkit that builds command-line interfaces so AI agents can control existing software for developers and power users.

  • skills

    skills is a command-line tool that finds, installs, and manages instruction packs that teach coding agents specific tasks for developers.

  • Claude Code Templates

    Claude Code Templates is a command-line tool that adds ready-made agents, commands, and settings to Claude Code for developers.

  • InsForge

    InsForge is a backend platform that gives coding agents a database, auth, storage, and hosting for building full-stack apps.

  • Inspector

    Inspector is a developer tool that tests and debugs MCP servers through a web UI, CLI, or terminal UI for developers building agent connectors.

  • Gentle-AI

    Gentle-AI is a configuration layer that adds memory, a workflow, and review steps to AI coding agents for developers working on projects.

More like this:

Last updated Nov 2, 2026. Tracked since Oct 28, 2026. The summary was written by an AI model from the project's own description and checked against our copy rules. Spotted an error? Tell us.