Getting started

Quickstart

Connect your first repository and run an AI-selected test pipeline in under 5 minutes.

  1. Create your account

    Go to superplane.com and sign up with GitHub OAuth or work email. GitHub OAuth is faster — it pre-authorizes repository access for the next step. Developer plan is free, no card required.

    Note — plan selection
    # Developer plan: free, no card required
    # Builder plan: $89/mo — unlocks env provisioning + canary rollout
    # Platform plan: $299/mo — adds SSO, audit log, RBAC, 99.9% SLA
  2. Connect a repository

    In the SuperPlane dashboard, click Add repository. Select the repo you want to connect. SuperPlane requests minimum required scopes only.

    GitHub OAuth permission request
    Requested scopes:
      repo:read        # read commit metadata + file tree
      checks:write     # post check run status
      statuses:write   # update commit status
  3. Create policy.yml

    Add a .superplane/policy.yml file to your repository root. This is the control plane for SuperPlane's behavior on each run.

    .superplane/policy.yml
    version: "1"
    
    test_selection:
      enabled: true
      confidence_threshold: 0.85
    
    env_provisioning:
      enabled: true
      provider: "aws"
    
    canary:
      enabled: false
  4. Add SuperPlane to your CI config

    Add a single step to your existing CI workflow. SuperPlane wraps your existing test runner — it doesn't replace it.

    .github/workflows/ci.yml
    steps:
      # ... your existing checkout and setup steps
    
      - name: SuperPlane — AI test selection
        uses: superplane-io/action@v1
        with:
          api_token: ${{ secrets.SUPERPLANE_TOKEN }}
          policy: ".superplane/policy.yml"
    
      - name: Run selected tests
        run: ${{ steps.superplane.outputs.test_command }}
  5. Push and review your first run

    Push a commit. SuperPlane intercepts the pipeline, runs AI selection, and posts the run log back to GitHub. You'll see which tests were selected, which were skipped, and the confidence score for each decision.

    SuperPlane run log — first run
    ✓ superplane v0.9.2
    ────────────────────────────────────────────
    commit        a7f3c21  feat: add export to CSV
    files_changed 3
    tests_total   847
    tests_selected 112  # 735 skipped at confidence >0.85
    ────────────────────────────────────────────
    api/export_test.py        SELECTED  0.97
    api/auth_test.py          SELECTED  0.91
    ui/dashboard_test.py      skipped   0.23
    tests/regression/suite_4  skipped   0.11

Next steps

The model starts with zero history — it gets smarter with every run. After ~20 runs you'll see selection accuracy climb significantly.