Getting started

Start the dashboard, put a control plane in your cloud, connect a repository and run your first job on your own machines.

You need an account with at least one of AWS, Cloudflare or Modal, and a repository on GitHub or GitLab whose jobs you want to run there.

1. Start the dashboard

npx @superci/cli dashboard

Without Node, install the same program as superci, then run superci dashboard:

brew install supercorp-ai/tap/superci           # Homebrew
curl -fsSL https://superci.dev/install.sh | sh  # macOS and Linux, into ~/.local/bin
cargo install superci                           # built from source by Rust

Your dashboard opens in the browser, on your own computer. You sign in with your cloud once: SuperCI keeps that sign-in in its own folder (~/.superci), and superci logout removes it. An AWS sign-in ends after twelve hours at most, as AWS has it.

2. Put a control plane in your cloud

Sign in with Cloudflare, AWS or Modal. The dashboard offers to deploy a control plane there, and shows what that creates before you press Deploy.

The control plane is small: a Cloudflare Worker, an AWS Lambda function, or a Modal app. It is the address GitHub and GitLab send job events to, and it starts a machine for each job. One control plane is enough, whichever clouds your jobs run in.

3. Connect your code

Under Repositories, connect GitHub or GitLab.

  • GitHub: the dashboard creates a GitHub App for you and you install it on the organization or the repositories you choose. Nothing is shared with a third party: the App is yours. GitHub lets only an organization’s owners create its App. If you are not one, the dialog has a note to copy and send to an owner. See GitHub.
  • GitLab: you create an access token and paste it in. See GitLab CI.

4. Add a runner provider

Under Runners, add where jobs should run: AWS, Cloudflare, Modal, or several. With more than one, you set their order. See Runner providers.

5. Change one line

In a workflow, replace the runner:

jobs:
  test:
    runs-on: superci
    steps:
      - uses: actions/checkout@v5
      - run: npm test

Push it. The job appears under Jobs in the dashboard, with the machine it got and what it cost.

To move a whole repository at once, Workflows has a prompt you can hand to a coding agent: it changes every workflow SuperCI can run and leaves the rest, saying why.

What next

  • Ask for a larger machine, a GPU or Windows in the label: Labels and machines.
  • Cap what a provider may run or spend: Order and limits.
  • The dashboard can be closed. Jobs keep running: the control plane does the work.