From zero to your first runner.

How Pi Dash connects projects, runners and agents, and how to install it on your own machines.

Three parts, one loop

The cloud

Hosted projects, issues and boards: where work is scoped, handed to agents and reviewed to done.

The command line

The local CLI that signs a machine in and carries commands, context and results both ways.

The runner

One agent bound to one project on a machine you control: where delegated work actually runs.

Get running in four steps

  1. Step 1

    Create a project

    Sign in, create a workspace and a project, and add at least one well-scoped issue so the first run has something concrete to drive.

  2. Step 2

    Install the CLI

    One command per platform downloads the installer, verifies it and drops the CLI on the path.

    macOS and Linux

    curl --proto '=https' --tlsv1.2 -LsSf \
      https://github.com/The-AI-Republic/pi-dash/releases/latest/download/install.sh | sh

    Windows (PowerShell)

    irm https://github.com/The-AI-Republic/pi-dash/releases/latest/download/install.ps1 | iex

    Managed Windows deployments can use the MSI installer instead. MSI installer

    A Linux-subsystem install only drives agents inside that subsystem; when the coding agents live on the Windows side, use the PowerShell command instead.

  3. Step 3

    Sign the machine in

    Device login pairs the host with the cloud. The installers start this step for you; run it by hand after an MSI install, on a headless host, or when re-authenticating.

    Device login

    pidash auth login
  4. Step 4

    Enroll a runner

    Pick the project and the agent in the product, then run the enrollment commands on the host and check the diagnostics.

    Enrollment

    pidash runner add --project <PROJECT_ID>
    pidash doctor

Give coding sessions the same access

The agent skill is optional: one command gives coding sessions product access through the CLI you already signed in.

Agent skill

npx @airepublic/pidash-skill-installer

The operating loop

  • DelegateScope an issue and hand it to a runner.
  • ExecuteThe agent implements on its own machine.
  • ValidateThe project's checks prove the change.
  • ReviewA second pass takes it to done.

Built to trust

Stays on your hostSource checkouts and credentials never leave the runner machine.
Access you grantMachine access is explicit, scoped and revocable at any time.
Open at the coreThe core is open source and self-hostable; the cloud is optional.