Opensbx

Opensbx is a self-hosted, API-first sandbox runtime for executing untrusted or AI-generated code in isolated Docker environments.

It is built for teams that want infrastructure control, predictable costs, and clear runtime boundaries.

Core Capabilities

• Create disposable sandboxes from Docker images.
• Execute commands and stream logs.
• Read, write, list, and delete files inside sandboxes.
• Expose sandbox services through generated subdomain URLs.
• Apply resource and timeout limits.
• Stop and delete sandboxes cleanly.

Why Teams Use Opensbx

• Self-hosted control: run on your own servers and network.
• Lightweight architecture: no Kubernetes control plane required.
• API-first workflows: easy integration with apps, tools, and agents.
• AI-ready runtime: designed for generated script and code execution.
• MCP support: includes a /v1/mcp endpoint.
• Low lock-in: you control images and operations.

Typical Use Cases

• AI coding agents and tool execution.
• Code execution products.
• User-submitted script runners.
• Disposable CI or test workloads.
• Internal dev and prototyping environments.

Conceptual Flow

1. Use an existing image like node:22 or python:3.12.
2. Create a sandbox with optional ports, env vars, timeout, and resource limits.
3. Run commands and read logs via the API.
4. Manage files and directories via the API.
5. Access exposed services through generated subdomain URLs.
6. Stop or delete the sandbox, or let timeout auto-stop it.

Next Steps