Labor0 Docs

Use the Labor0 CLI

The Labor0 CLI lets you run eligible project work through a local runner on a computer you control. It is for developers and workspace operators who need a repeatable terminal setup, including unattended or login-service use.

Supported platforms

The CLI supports macOS 11 or later, Windows 10 or later, and Linux kernel 4.1 or later. The login service additionally needs macOS 13 or later, Windows 10 build 1809 or later, or Linux with systemd 252 or later and a working user service manager.

Before you begin

Have access to an eligible Labor0 workspace, permission to run work there, a supported coding agent (Codex, Claude Code, or OpenCode), and a small task that does not contain secrets or private customer data. The CLI needs network access to Labor0; it does not install or sign in to your coding agent.

Shortest safe path

  1. Install and verify the complete bundle.
  2. Authorize one workspace and start the runner.
  3. Manage, recover, or hand off support safely.
  4. Use the complete command reference for product and administration workflows.

For a privacy-safe first task, use an input such as: “In a scratch branch, add a README heading named Local runner check and show the changed file.” In Labor0, confirm the task detail shows the intended workspace and final Local runner attribution before allowing work to continue. The visible completion evidence is the changed file in the review result; stop before retrying if the attribution is not Local runner or if hosted fallback is unacceptable.

Continue with Run your first task, then review Pricing and usage for eligibility and cost terms.

Choose a guide

  • Install the CLI — choose the right archive, verify SHA-256, extract the full bundle, set PATH, recover platform warnings, and update safely.
  • Set up and run — check Codex, Claude Code, or OpenCode, run l0 doctor, authorize a workspace, give consent, start/status the runner, and complete a bounded first task.
  • Manage and troubleshoot — select workspaces, operate the runner or login service, inspect or clear cache, create safe support evidence, and recover failures.
  • Command reference — authenticate, manage scope, connect GitHub, operate Graphs, tasks and sessions, and administer access, security, and billing.
  • Run your first task — choose Auto or understand the current Local only restriction before submitting work.
  • Pricing and usage — review local-execution eligibility, Sandbox terms, and customer-paid provider costs.

Browser execution boundary

In the browser, Auto prefers an eligible active local runner but may use hosted fallback when local execution is unavailable and hosted execution is allowed. An active online CLI-managed runner does not currently enable Local only because Labor0 cannot confirm readiness for the selected coding-agent runtime. Local only never falls back to hosted execution. Before approving a task, confirm the final Local runner attribution in task or session details. If hosted fallback is not acceptable, do not submit the browser task with Auto, because hosted work may begin before final attribution is available; stop before retrying.

What ready looks like

The selected CLI guide is complete when its visible command result, Labor0 workspace selection, and final Local runner attribution all agree.