BoltzLabs
sign in sign up
  • Overview
  • CLI quickstart
  • Desktop app quickstart
  • Python SDK quickstart
  • JavaScript / TypeScript quickstart
OverviewCLI quickstartDesktop app quickstartPython SDK quickstartJavaScript / TypeScript quickstartCLIDesktop appPython SDKJavaScript / TypeScript SDKCLIDesktop appPython SDKJavaScript / TypeScript SDKHTTP APISupported languagesOverviewRun an evalUse your own modelsResults and tracesTask setsPython tracerSandboxesRun codeEvalsUsage and limits

Get started

Documentation

Choose how you want to get started.

CLI

Work from your terminal.

Set up the CLI →

Desktop app

Use sandboxes and Run code from a Mac app.

Install the desktop app →

Python

Use sandboxes and Run code from Python.

Start with Python →

JavaScript / TypeScript

Use sandboxes and Run code from Node.js.

Start with JavaScript / TypeScript →

What you can run

Sandboxes

Isolated machines with a shell, files and ports that stay up while you work.

Run code

Send code and a language, get the output back. No sandbox to create or clean up.

Traces

See every step your agent took. Add two lines to a Python agent, or compare models on the same task.

Before you start

Sign in and create an API key in your dashboard. For access, usage, and account limits, see Account limits.

next CLI quickstart
BoltzLabs

Disposable sandboxes for agents, LLMs, and RL.

Product

  • Docs
  • Explore
  • Solutions
  • Blogs
  • Pricing
  • Contact

Developers

  • CLI
  • Python SDK
  • Python tracer
  • TypeScript SDK

Use cases

  • Coding agents
  • LLM execution
  • RL rollouts
  • Evals
© 2026 BoltzLabs