> ## Documentation Index
> Fetch the complete documentation index at: https://docs.slipway.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Slipway

> Preview environments for every push and pull request. Ship from a manifest.

Slipway watches your GitHub repositories and brings a real, isolated, internet-reachable copy of your app up on every push to the default branch and on every open pull request. You write one `slipway.yaml`, and slipway builds the images, gives each service its own URL, and tears everything down on its own when the PR closes or the TTL expires. It's the same environment your team reviews in, and one your AI agents can drive over MCP.

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Install the GitHub App, add a `slipway.yaml`, and get a live preview URL in under five minutes.
  </Card>

  <Card title="Core concepts" icon="compass" href="/concepts">
    Deployments, services, ephemeral previews, supersession. The shape of how slipway thinks.
  </Card>

  <Card title="Slipway spec" icon="file-code" href="/slipway-yaml/overview">
    Every shipped field of the manifest, with worked examples and a full field index.
  </Card>

  <Card title="AI agents" icon="robot" href="/ai-agents">
    Hand Claude, Cursor, or your own agent a real place to run. Slipway ships a first-party MCP server.
  </Card>

  <Card title="Deployments" icon="code-pull-request" href="/deployments/overview">
    Triggers, manual deploys, and what happens at each step of the lifecycle.
  </Card>

  <Card title="Environments" icon="layer-group" href="/environments/overview">
    Multi-repo stacks, PR previews, and the lifecycle from create to teardown.
  </Card>
</CardGroup>

<Frame caption="A live PR preview. Every push to the PR rebuilds; closing the PR tears the preview down automatically.">
  <img src="https://mintcdn.com/slipway/XyuXtxn4cFnSJANr/images/deployments/detail.png?fit=max&auto=format&n=XyuXtxn4cFnSJANr&q=85&s=323285c6e0438f451a3d2b41336465a9" alt="Deployment detail page" width="1920" height="1200" data-path="images/deployments/detail.png" />
</Frame>

## What it's for

* **PR previews.** Every pull request gets its own live environment at a stable URL (`<env-name>-<service>-<id>.<your-apps-domain>`). Teardown is automatic on PR close.
* **Default-branch previews.** Every push to `main` rebuilds and supersedes the previous environment.
* **Ephemeral one-offs.** Spin up any ref as its own isolated instance with `sw dev up`, on its own URL, reaped automatically once you're done.
* **Built for AI agents.** slipway ships a first-party MCP server. Point Claude, Cursor, or your own agent at it and it drives the whole lifecycle: bring a branch up to a live URL, read the logs, exec into a container to verify a fix, then tear it down. The verification loop agents otherwise lack. See [AI agents](/ai-agents).

Slipway is a **dev tool**. Every deployment is ephemeral by design. Don't ship production traffic to a slipway preview; ship production from somewhere else.

## How it works

```mermaid theme={"system"}
flowchart LR
  A[GitHub push / PR] --> B[Slipway]
  B --> C[Build images]
  C --> D[Start services]
  D --> E[Live URL]
```

Every deployment runs in its own isolated environment. Other tenants on the platform can't reach your services, and your services can't reach theirs.