# Quickstart (https://luxor.ai/docs/runners/quickstart)

> For the complete documentation index, see [llms.txt](https://luxor.ai/llms.txt)

Install the Luxor Agent Cloud GitHub App, switch one line in your workflow, and your jobs run on Runners.
Caching works out of the box.

Runners are a drop-in replacement for GitHub-hosted runners: the only workflow change is the `runs-on` value. This guide takes you from sign-up to your first job on Runners in a few minutes.

**Step 1:**
## Connect your GitHub organization

[Create an account](https://app.tenki.cloud/auth/registration/) (email or GitHub), then open the [dashboard](https://app.tenki.cloud), select **Runners**, and start the GitHub connection. On GitHub, choose which repositories the **Luxor Agent Cloud** app can access and select **Install & Authorize**:

<img src="getAsset(&#x22;/images/get-started/github-install.png&#x22;)" alt="GitHub installation page for selecting Runners repository access" />

**Step 2:**
## Switch your workflows to Runners

The fastest path is the **Migration Wizard**: in the Runners dashboard, select a repository and it detects every workflow's current runner, lets you pick a runner size per job, and opens a pull request with the changes. Merge it and your jobs run on Runners.

[Video](https://storage.googleapis.com/tenki-cloud-assets/docs/tenki_step3_migration_wizard.mp4)

The wizard never touches anything other than `runs-on`. Workflows it can't rewrite safely (YAML anchors, multi-line `runs-on` values) are skipped and listed in the result so you can edit them by hand.

To migrate manually instead, replace the runner tag in `.github/workflows/*.yml`:

```yaml
jobs:
  build:
-    runs-on: ubuntu-latest
+    runs-on: tenki-standard-medium-4c-8g
```

Migrating from self-hosted runners? Replace the entire `runs-on` value—including any label array—with one runner label.

See [Runner Sizes & Labels](https://luxor.ai/docs/runners/sizes.md) for available labels, or [Which runner should I use?](https://luxor.ai/docs/runners/which-runner-should-i-use.md) for sizing guidance.

**Step 3:**
## Run your first job

Merge the Migration Wizard pull request—or commit your manual edit—then trigger the workflow as you normally do. In GitHub's **Actions** tab, open the run and confirm the job completes on the runner label you selected.

Existing `actions/cache` steps keep working unchanged, and package downloads from npm, apt, Go modules, Maven, and other registries are cached automatically. See [Caching](https://luxor.ai/docs/runners/caching.md) for quotas and details.

***

## Next steps

* [Runner Sizes & Labels](https://luxor.ai/docs/runners/sizes.md): standard labels and specifications
* [Caching](https://luxor.ai/docs/runners/caching.md): the hosted Actions cache and dependency mirrors
* [Sticky Disks](https://luxor.ai/docs/runners/sticky-disks.md): persistent cache disks that stay warm between jobs
* [Limits & Concurrency](https://luxor.ai/docs/runners/limits-and-concurrency.md): what to expect at peak load
* [Networking & Egress](https://luxor.ai/docs/runners/networking.md): egress IPs and private connectivity
* [Security & Isolation](https://luxor.ai/docs/trust/security.md): ephemeral VMs and the isolation model

Hit a snag, like your organization not showing up or jobs stuck in queue? See [Troubleshooting](https://luxor.ai/docs/runners/troubleshooting.md).