Verified project record

opensandbox-group/OpenSandbox

OpenSandbox is a secure, fast, and extensible general-purpose sandbox platform for AI applications. It provides unified APIs, multi-language SDKs, and Docker/Kubernetes runtimes for coding agents, GUI agents, and agent evaluation.

Project overview

It solves the core challenge of securely executing AI-generated code and commands by combining strong isolation technologies with a multi-runtime architecture that scales from local development to large-scale distributed scheduling.

Project type: MCP · AI Agent · Infrastructure
Use cases: Coding & Development · Automation
Deployment: Refer to project documentation
License: Apache-2.0

Best for

Developers and AI engineers who need a secure, extensible sandbox platform for running AI agent code and commands with strong isolation guarantees.
Operations teams requiring large-scale distributed scheduling of isolated sandbox workloads across Docker or Kubernetes environments.

Key capabilities

Built-in lifecycle management supporting Docker and high-performance Kubernetes runtime, enabling both local runs and large-scale distributed scheduling.
Provides SDKs for Python, Java/Kotlin, JavaScript/TypeScript, C#/.NET, and Go for sandbox lifecycle, command execution, and file operations.
Terminal CLI for the common sandbox workflow: create sandboxes, run commands, move files, inspect diagnostics, and manage runtime egress policy.
Unified ingress gateway with multiple routing strategies plus per-sandbox egress controls.
Secure credential injection for sandbox outbound requests without exposing real secrets to workloads.
Supports secure container runtimes like gVisor, Kata Containers, and Firecracker microVM for enhanced isolation between sandbox workloads and the host.

Limitations and risks

Docker must be installed and available for local sandbox execution.
Python 3.10 or higher is required for running examples and the local runtime.

Getting started

Install Docker and Python 3.10+, then run uvx opensandbox-server init-config ~/.sandbox.toml --example docker to generate configuration. Start the server with uvx opensandbox-server. Create your first sandbox using osb sandbox create --image python:3.12 --timeout 30m -o json.

Evidence and sources

README: OpenSandbox is a **general-purpose sandbox platform** for AI applications, offering multi-language SDKs, unified sandbox APIs, and Docker/Kubernetes runtimes for scenarios like Co…
README: - 🚀 **Sandbox Runtime**: Built-in lifecycle management supporting Docker and high-performance Kubernetes runtime, enabling both local runs and large-scale distributed scheduling.…
README: Python: ```bash pip install opensandbox ``` Java/Kotlin (Gradle Kotlin DSL): ```kotlin dependencies { implementation("com.alibaba.opensandbox:sandbox:{latest_version}") } ``` Java…
README: OpenSandbox also provides `osb`, a terminal CLI for the common sandbox workflow: create sandboxes, run commands, move files, inspect diagnostics, and manage runtime egress policy.
README: Create a sandbox and execute commands and codes.

Reviewed at: 2026-07-21T22:28:33Z