
We build sites on both. We maintain two shop themes on Astro and EmDash: Kai, where EmDash runs the whole shop, and Nalu, an Astro storefront in front of an existing WooCommerce store. This is what changes when a small shop leaves WordPress, from the developer’s side. It is not a verdict on WordPress; it is a list of what you will be doing differently. For the business side of the question, without code, our agency site explains moving from WordPress to Astro and EmDash and what each option costs.
Facts about EmDash come from its documentation and the 1.0 announcement. Facts about our themes come from their code.
1. There is no shop plugin to install, yet
This is the first thing to know. EmDash is a CMS, not a commerce platform. Its documentation page for WordPress developers maps posts, custom post types, taxonomies, menus and widgets, and says nothing about WooCommerce. The plugin registry that launched with 1.0, on 28 September 2026, supports free plugins; paid plugins are planned for later. The same announcement mentions a third-party eCommerce plugin in preparation.
So today, a shop on EmDash means one of three things:
- a commerce layer someone wrote for EmDash, which you evaluate like any young dependency;
- your own code;
- keeping the commerce engine you already have and changing only the storefront.
Kai is the second option, shipped as source. Nalu is the third.
2. The theme is the site
On WordPress, a shop theme is a skin: WooCommerce owns the templates, the theme overrides some of them, and you can switch themes. On EmDash, the theme is a complete Astro project. There is no runtime theme package and no template hierarchy. Once you start from it, the files are yours.
For a shop, that means the product page, the shelf pages and the cart are ordinary Astro routes and components that you read and change directly. In Kai: a product collection validated at build time, a filterable and sortable shop grid, product pages with options and stock, shelf pages, and a cart kept in the browser behind one typed module.
The cost: when you want a feature, there is no theme option to tick. You write it, or the theme author did.
3. Plugins: two formats instead of one
WordPress has one kind of plugin, running with full access. EmDash has two:
- sandboxed, installed from the registry, running in a separate runtime and limited to the APIs it declares;
- native, installed from npm through the Astro configuration, running in the same process as the site, able to add React pages to the admin. It needs a deploy.
A shop back office (orders, refunds, stock, invoices) needs admin screens and database tables, so in practice it is native code. In Kai, the shop screens are React files inside the theme, alongside the stock module. You can read every line, and you also carry them: at each EmDash release we check them against the new admin. For 1.5.0 that work is described in our upgrade notes.
4. Payment and the rest are configuration, not marketplaces
What Kai does with the engine on: Stripe orders through hosted Checkout, refunds, stock counted per product and per option (a paid order counts it down, a sold-out product can no longer be bought, an alert email warns you), shipments, invoices, prices shown with or without VAT.
The other services are plugins delivered with our themes, where the key of the service is pasted in the back office: Stripe, PayPal, SumUp, Brevo, Colissimo, Mondial Relay. They are listed on our EmDash plugins page; Pennylane invoicing comes with the Kailua 3 Pro licence. When a shop takes payment through SumUp or PayPal and has no Stripe key, the Stripe fields leave the product form; the values stay in the database.
What you do not get is a marketplace of extensions for every edge case. If your shop depends on a specific WooCommerce extension, list those dependencies first. Each one is either rewritten, replaced by a service with an API, or a reason to keep WooCommerce.
5. Keep WooCommerce, change the storefront
That last case is what Nalu is for. The WooCommerce store keeps running: orders, payment, stock and the WordPress admin stay where they are. Astro becomes the storefront. One URL to fill in. With the URL empty, the theme runs on its own demo catalogue, which is what the online demo shows. A shop connected to WooCommerce keeps WooCommerce’s stock.
This is the lowest-risk move: nothing to migrate on the commerce side, and the public pages get the Astro build. The WordPress install is still there to maintain.
6. Hosting and data
EmDash runs on Cloudflare (Workers, D1, R2) or on a Node.js server with SQLite. Our themes use Cloudflare when the engine is on. The shop data lives in a D1 database in the buyer’s own Cloudflare account.
Two things change for the person who runs it:
- Backups.
wrangler d1 exportdoes not support databases with virtual tables, and EmDash search uses FTS5 tables, which are virtual (Cloudflare documentation). On our sites the export fails. The recovery point is D1 Time Travel, which keeps 30 days on the paid plan and 7 on the free one (Time Travel). Plan it before you sell anything. - No PHP, no server to patch. The site is an Astro build deployed as a Worker. The work moves from updating a server and plugins to updating packages and redeploying.
7. Most of the shop can stay static
Kai builds two ways from one repository. Engine off, pnpm build writes plain HTML: catalogue, product pages, shelves, cart. The shop filter and sort reorder items the server already rendered, and the theme ships no React on the public site. The checkout button is then disabled with a sentence explaining that no payment provider is wired.
Engine on, the same pages are rendered by a Worker on EmDash, and the shop module replaces that button with Stripe Checkout once the two Stripe secrets are set. So a shop can start as a fast static catalogue and switch on the back office when it is ready to take orders.
8. Migration moves content, not the shop
EmDash has a WordPress importer, and 1.5.0 added an emdash import wordpress command that converts a WXR export into files on disk, with a redirect map. The migration documentation lists what comes across: posts, pages, custom post types, taxonomy terms, menus, reusable blocks, authors and media.
The theme does not come across: templates have to be rebuilt as Astro routes. And the migration page does not mention WooCommerce products, orders or customers. For a shop, plan the catalogue import as its own task, and decide what happens to order history.
What changes, in one list
- No commerce platform included: you choose a commerce layer, your own code, or a headless storefront.
- The theme is your source code, not a switchable skin.
- Shop admin screens are native code you own and maintain at each EmDash release.
- Payment and services are configured, not picked from a marketplace.
- Data lives in your D1 database; backups go through Time Travel.
- The public shop can be fully static until you take orders.
- Content migrates; the shop logic and the order history are a project.
If those trade-offs fit, the two themes cover both paths: Kai for a shop run from EmDash, Nalu to keep WooCommerce behind an Astro storefront. Both demos are online, and the documentation shows how to connect the sale. The three ways to sell with EmDash are compared in building an ecommerce site with EmDash.



