Skip to main content
Migration: The CLI is now @isol8/cli (installed as isol8 command). The library is @isol8/core.
This quickstart is intentionally practical. You will:
  1. install and verify the CLI
  2. build runtime images
  3. run real sandboxed code
  4. check effective defaults

​
Prerequisites

  • Docker installed and running
  • Bun or Node.js runtime available
Default behavior: network: "none", timeoutMs: 30000, memoryLimit: "512m".
1

Install and verify

Install the global CLI.
npm install -g @isol8/cli
Verify installation and Docker connectivity:
isol8 --version && docker info >/dev/null
2

Build runtime images

Run setup to build the base Docker images (isol8:python, isol8:node, etc.). This creates the local isol8 image hierarchy.
isol8 setup
This step is idempotent. Is uses Docker labels to detect if images are already up-to-date, so you can run it safely as part of your deployment script.
3

Run code

Execute a simple Python script.
isol8 run -e "print('hello from sandbox')" --runtime python
4

Try core capabilities

Install a package (ephemeral) and inspect config.
# Install numpy on the fly
isol8 run \
  -e "import numpy; print(numpy.array([1,2,3]))" \
  --runtime python \
  --install numpy

# Check resolved configuration
isol8 config

​
Next steps

Execution guide

Learn request fields, modes, and streaming behavior.

CLI reference

Complete command and flag behavior.

Option mapping

Map every feature to CLI/config/API/library inputs.

Troubleshooting

Diagnose setup, runtime, and server issues.