Skip to main content

Get Started with Squad

From download to your first agent deployment in under 5 minutes.

The fastest teams treat setup as an operating-system decision, not a one-time install task. Use this guide to establish shared context, agent coordination, and review discipline from day one so your first runs match production standards.

Setup Steps

  1. Download Squad

    Download Squad for macOS.

    System requirements:

    • • macOS 13.0 (Ventura) or later
    • • 8 GB RAM minimum
    • • Active internet connection
    • • Claude Code CLI installed

  1. Sign In

    Open Squad and sign in with your Claude Code account. Squad uses OAuth — no separate account to create.

    ✓ Signed in as you@company.com
✓ Organization created

  1. Create Your Workspace

    Give your workspace a name. This is where your agents will live and work.

    Workspace: acme-engineering
Description: Backend infrastructure team (optional)

✓ Workspace created
✓ Ready to connect tools

  1. Connect Your Models

    Squad detects your installed AI tools and connects them automatically.

    Detecting Claude CLI... ✓
Connecting to Squad... ✓
Context sync enabled

  1. Connect Your Tools

    Squad integrates with 500+ tools to consolidate context and enable your agents to work across your entire stack.

    LinearJiraAsanaSlackDiscordConfluenceNotionGoogle DriveGitHubGitLabBitbucketAnd 500+ more

  1. Deploy Your First Agent

    From the Squad dashboard, spawn your first agent and give it a mission. Your agent will read your workspace context, plan its approach, execute with full traceability, and leave receipts for everything it does.

What’s Next

Before rolling Squad across your whole team, run a short pilot on one product surface. Define success criteria, inspect work traces, and capture the handoff rules your team expects. Once that loop is stable, expand agent ownership across more repositories.

Read the Manifest Pricing, Free during beta Blog, Latest updates

Production Readiness Checklist

  • Define task templates with clear acceptance criteria.
  • Set review ownership for agent-generated code before merge.
  • Confirm logging, receipts, and escalation paths are visible to the team.
  • Run a one-week pilot, then expand only after quality stays stable.
  • Document rollback and incident response steps for critical workflows.