Why use sandboxes
Traditional containers share the host kernel, which means a container escape vulnerability can compromise the entire system. Sandboxes solve this by running each container in its own microVM with a separate kernel instance.
Use sandboxes when you need to run code you don't control:
- LLM-generated code: Execute code written by AI models without risking your infrastructure
- User-submitted code: Run code playgrounds, online IDEs, or code evaluation platforms
- AI agents: Provide isolated execution environments for autonomous agents
- CI/CD pipelines: Run builds and tests in secure, ephemeral environments
- Multi-tenant platforms: Isolate customer workloads with VM-level security boundaries
How sandboxes work
Sandboxes use a combination of microVM-based virtualization and user-space kernel isolation to provide strong isolation. Depending on the technology, each container runs either in a lightweight virtual machine with its own kernel or in a dedicated user-space kernel sandbox, preventing breakout attacks while maintaining near-native performance.
Deploy sandboxes
Choose where to run your sandboxes based on your requirements.
Programmatic control
Automate sandbox provisioning and lifecycle management with the Northflank API and SDK.