Skip to content
Use this GitHub action with your project
Add this Action to an existing workflow or create a new one
View on Marketplace

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

48 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

shipstatic/action

GitHub Action for ShipStatic — deploy static websites, landing pages, and prototypes instantly from CI.

Deploy — Free, No Account Needed

name: Deploy
on: push

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - run: npm ci && npm run build
      - uses: shipstatic/action@v2
        with:
          path: ./dist

Your site is live instantly on *.shipstatic.com. No token, no sign-up, no configuration.

Deployments made without a token are public and expire — the claim output holds a URL that keeps the site permanently, and expires holds the deadline.

On pull requests the action comments the URL automatically. Give the workflow pull-requests: write and there is nothing else to configure.

All Features — Free API Key

For permanent deployments and full control over your sites and domains, get a free API key from my.shipstatic.com/api-key:

  1. Get a free key at my.shipstatic.com/api-key
  2. Add it as a repository secret named SHIP_TOKEN (Settings > Secrets and variables > Actions)
name: Deploy
on:
  push:
    branches: [main]
  pull_request:

permissions:
  contents: read
  pull-requests: write

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - run: npm ci && npm run build
      - uses: shipstatic/action@v2
        with:
          token: ${{ secrets.SHIP_TOKEN }}
          path: ./dist
          domain: www.example.com

domain deploys and links in one step: when the action finishes, your site is live at that address and the url output is your domain's URL. It needs a token — a domain belongs to an account — and without one the deploy is refused before anything is uploaded, rather than silently landing somewhere else.

One slot, two kinds of token

token takes either credential ShipStatic issues:

Value What it is
ship-… An API key — durable, full account access
deploy-… A deploy token — scoped to deploys, optionally time-limited, revocable

A deploy-only workflow can run on a deploy token. This action never reads your account, so nothing here needs the wider credential — create one at my.shipstatic.com and keep the API key out of CI.

Inputs

Input Required Default Description
token No Your API key (ship-…), or a scoped deploy token (deploy-…) — one slot takes either. Omit to deploy anonymously
api-url No production API ShipStatic API endpoint
path No . Directory to deploy
domain No Domain to serve the deployment at — deployed and linked in one step (requires token)
ttl No Seconds until the deployment expires and is cleaned up (requires token; cannot be combined with domain)
password No Password-protect the deployment (6–128 characters). Visitors are prompted to unlock before viewing
labels No Comma-separated labels, added to the automatic commit label
idempotency-key No derived Override the replay key (see below)
github-token No the workflow's own token GitHub token for the PR comment. Pass an empty string to post no comment

github-token enables one thing: a PR comment carrying the deployment URL, posted once and updated in place on every push. It defaults to your workflow's own token, so the only thing to configure is the permission to write with it:

permissions:
  contents: read
  pull-requests: write

Pass github-token: '' if you would rather the action never commented. Pass a different token to comment on a PR in another repository.

To record deployments in your repo's sidebar, declare an environment: on the job. GitHub writes that record itself, with the environment you named — this action does not write one.

Outputs

Output Description
deployment Deployment hostname (e.g. happy-cat-abc1234.shipstatic.com)
url URL of the deployed site — your domain's URL when domain is set, otherwise the deployment's own (e.g. https://happy-cat-abc1234.shipstatic.com)
claim Claim URL — anonymous deployments only (visit to keep permanently)
expires Expiry as a unix timestamp in seconds — set for anonymous deployments and for any deployment given a ttl
      - uses: shipstatic/action@v2
        id: deploy
        with:
          path: ./dist
      - run: echo "Live at ${{ steps.deploy.outputs.url }}"

Every deploy also writes a summary table to the workflow run page — the deployment and its URL, plus the claim link and the expiry where the deployment has them.

Labels

Every deployment is labelled with the commit's short SHA automatically. labels adds your own on top:

      - uses: shipstatic/action@v2
        with:
          token: ${{ secrets.SHIP_TOKEN }}
          path: ./dist
          labels: preview,pr-${{ github.event.number }}

Labels are 3–25 characters, lowercase alphanumeric with ., _ or - between segments, up to 10 per deployment. Find them again with ship deployments list.

Previews that clean themselves up

ttl gives a deployment a lifetime in seconds. When it runs out the platform reclaims the deployment — no cleanup workflow, no pull_request: closed handler to maintain:

name: Preview
on: pull_request

permissions:
  contents: read
  pull-requests: write

jobs:
  preview:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - run: npm ci && npm run build
      - uses: shipstatic/action@v2
        with:
          token: ${{ secrets.SHIP_TOKEN }}
          path: ./dist
          ttl: 604800   # one week

ttl needs a token — a deployment made without one already expires on the platform's own schedule — and it cannot be combined with domain: a domain should not point at something that is about to be reclaimed. The expires output carries the deadline as a unix timestamp.

Re-running a job does not deploy twice

Every deploy carries an idempotency key derived from the workflow run:

<owner>/<repo>-<run id>-<job id>

A GitHub run id is stable across re-runs and fresh on every new push, so pressing Re-run jobs replays the deployment the first attempt created — same URL, no second upload, no second deployment in your account — while the next push deploys normally. Nothing to configure.

Two cases need an explicit key, because no ambient value distinguishes them:

  • Matrix legs deploying different content. All legs of a matrix share one job id.
  • Two deploys in one job. Both invocations share the run and the job.

Give each its own key:

    strategy:
      matrix:
        site: [docs, marketing]
    steps:
      - uses: shipstatic/action@v2
        with:
          path: ./dist/${{ matrix.site }}
          idempotency-key: ${{ github.run_id }}-${{ matrix.site }}

Deploying to a different endpoint

api-url points the action at another ShipStatic API. Omit it and the production API is used.

      - uses: shipstatic/action@v2
        with:
          token: ${{ secrets.SHIP_TOKEN }}
          api-url: ${{ vars.SHIP_API_URL }}
          path: ./dist

Requirements

GitHub-hosted runners need nothing. Self-hosted runners need Node.js 20 or newer and jq on PATH; the action installs the ShipStatic CLI itself.

Versioning

@v2 is a moving tag that always points at the latest 2.x release, and it speaks the ShipStatic 2.x platform — the action's major and the platform major it targets are the same number, by design. Pin an exact @vX.Y.Z tag for an immutable reference, or @<full-sha> if your supply-chain policy requires actions pinned by commit.

@v1 is frozen and speaks the 1.x platform. It receives no changes.

Things this action deliberately does not have

  • No cli-version input. The action's major names the CLI major it speaks; letting a workflow float across that boundary is the failure this design exists to prevent.
  • No SPA / path-detection toggles. The CLI's defaults are right, and a repo that needs different behaviour ships a ship.json, which the detection defers to.
  • No setup-node step. Hosted runners already carry a supported Node; installing another would slow every workflow for the exception's sake.
  • No timeout inputs. The CLI owns its own budgets, sized to the platform's upload limits.

Examples

Five copy-pasteable workflows in the action-example repo:

License

MIT