Runs AI-agent workloads in isolated Docker or Kubernetes sandboxes, exposing sandbox lifecycle, command, filesystem, and code-interpreter APIs across five language SDKs.
Use it now
Keep it with up to 9 other tools you're setting up, then share them all as one link.
Try this first
Ask your assistant, with OpenSandbox connected:
- Sandbox Lifecycle Management
Use OpenSandbox to create container sandboxes: ‹what you need›
- Code Interpreter
Use OpenSandbox to execute Python and other supported languages inside a sandbox: ‹what you need›
- Command And File Operations
Use OpenSandbox to run shell commands: ‹what you need›
Connect it over MCP
Connect OpenSandbox as an MCP server, following its setup instructions at https://open-sandbox.ai
Its page: open-sandbox.ai
Install the CLI
Install it as its own instructions say.
Its page: open-sandbox.ai
Call the API
Auth An API key from OpenSandbox Group
Its page: open-sandbox.ai
Compare
Press Compare to add it to the tray at the bottom of the window, then pick up to 2 others anywhere on the site.
- Access
- APIMCPCLI
- Auth
- API Key
- Skills
- 5
Skills
Sandbox Lifecycle Management
Creates, times out, and kills container sandboxes from a chosen image through one lifecycle API on Docker or Kubernetes.
Code Interpreter
Executes Python and other supported languages inside a sandbox and returns the result value, stdout logs, and exit state.
Command And File Operations
Runs shell commands and reads or writes sandbox files from the SDKs, the osb CLI, or the MCP server used by coding agents.
Network Policy Control
Routes inbound sandbox traffic through a unified ingress gateway and applies per-sandbox egress rules to restrict outbound calls.
Credential Vault Injection
Injects credentials into sandbox outbound requests at the proxy layer so workloads never see the real secret values.