Skip to content

README.md

Latest commit

 

History

History
45 lines (29 loc) · 2.05 KB

README.md

File metadata and controls

45 lines (29 loc) · 2.05 KB
Raw

pages-seo-install

One-command CLI installer for pages-seo.

Usage

Run the installer straight from the live site (no npm publish needed) — pick the runtime you have:

curl -fsSL https://seo.benjaminb.xyz/install/run.sh | bash    # Bash / Zsh
curl -fsSL https://seo.benjaminb.xyz/install/run.py | python3 # Python
curl -fsSL https://seo.benjaminb.xyz/install/run.js | node    # Node

That's it. The script will:

  1. Check that wrangler is installed (offers to install it for you).
  2. Run wrangler login if you're not already logged in to Cloudflare (opens your browser; no token to copy).
  3. Prompt for your project slug, site name, admin email, and password.
  4. Provision a D1 database and R2 bucket on your Cloudflare account.
  5. Download the latest pages-seo source, patch wrangler.toml, and run wrangler pages deploy — uploading both the static assets and the Functions bundle.
  6. Set SITE_NAME and SITE_URL as Pages environment variables.
  7. Open your new site's /admin with the credentials baked into the URL hash, so the first-run setup card auto-creates your account.

Total time: ~2 minutes including the wrangler login.

Why a CLI?

Cloudflare's public REST API doesn't currently support uploading a Pages Functions bundle — that's the bit that powers the admin dashboard and the API in this project. The only documented path that supports Functions is wrangler pages deploy, which is a local tool. So this CLI is the honest, working version of "one-click install."

If Cloudflare ever ships a public Functions Direct-Upload API, we can move the whole flow to a browser installer and delete this package.

Requirements

  • Node 20 or newer
  • A Cloudflare account (free tier is fine)
  • macOS, Linux, or WSL on Windows (PowerShell hasn't been tested)

Re-running

The installer is idempotent. If something fails partway through, run it again with the same project slug — it'll skip the resources that already exist and pick up from the failed step.

Licence

MIT — see LICENCE.