Northwind has one coffee and needs eight, plus a couple more origins to hang them off. None of that teaches anything — it is the same authoring you did in part 4, seven times, with different tasting notes.
So this part hands it to you. Import the checkpoint bundle and the catalogue is there: eight coffees, five origins, seventeen Experience nodes, every photograph.
What is worth reading is why it is an import, because the obvious alternative would have been a script.
Why this is an import
It would be easy to ship a seed.mjs that writes the catalogue for you. We have one — it builds the reference organization, it is in the sample repository, and you are welcome to read it. This series still hands you a bundle instead, and the reason is worth a paragraph.
There is a public authoring API: the Content Management API, on Pro and above. A key is an ordinary bearer credential, minted in Content under Settings → API keys and bound permanently to one authoring environment and one organization role — so a script writing through it is governed by exactly the branch rules, workflow gates and permissions the person holding that role is. It is how ebitex.io’s own content model is applied. Content-as-code against it is a legitimate thing to build.
It is still the wrong tool for this step. Part 7 authors nothing — it moves a finished catalogue from our organization into yours, which is what import is for. A script that recreated the same eight coffees would be a longer road to the same rows, and what you would learn on the way is the script.
Authoring is a script’s job. Moving a finished thing between organizations is import’s.
And import is a real, shipped, supported feature with a screen and a permission. It is how content is meant to move between organizations, and a bundle was already how this sample distributes its checkpoints. There was nothing to invent.
What is in the bundle
step-07.zip carries the whole model as it stands: 9 Contracts, 5 Templates, 2 Category Groups with 8 categories, 3 folders, 15 Components and 17 Experience nodes — plus all 13 image blobs, as their own entries in the zip beside the manifest, so an imported copy has working pictures rather than dangling references to somebody else’s files.
Import it with fresh identity. That rewrites every id and every reference between them, so what lands is genuinely yours rather than a copy carrying our identifiers around. And an import writes drafts — nothing is delivered until you publish, which is part 2’s rule arriving in bulk.
What a bundle does not carry
This is worth knowing now rather than in part 15, because the failure mode is quiet.
A bundle carries content. It does not carry configuration — locales, workflow definitions, environments and delivery keys all belong to the organization and stay there.
For most of those the consequence is obvious: import into an organization with no workflow definitions and nothing is governed. Localization is the dangerous one, because it fails in the other direction. The translated text travels inside every document, exactly where it lives; what does not travel is the locale that selects it. So the site serves its default language, every page renders perfectly, and nothing anywhere reports a problem — the translations are simply never asked for.
If you import a bundle from a multilingual organization, add the locales before you go looking for the translations.
Where Starter stops
This is the first of exactly two places this series mentions a plan limit, and it is here because it is where the sample walks into one while doing something entirely reasonable.
Northwind is now seventeen Experience nodes. Starter allows ten. Parts 1 to 6 built seven of them and fit comfortably; this part crosses the line three pages into the catalogue, with a plain refusal:
403 quota_exceeded — content.max_experience_nodes
This organization has reached its plan’s Experience node limit (10). Upgrade to create more.
Two things about that are worth saying plainly, because they are the honest shape of it.
No Content capability is restricted by tier. Not localization, not personalization, not workflow, not Adapters, not streams, not contextual values. There is no feature flag and no quota key for any of them. Everything in this series except one step runs identically on either plan.
What a plan buys is scale and environments. Nodes, Components, sites, categories, storage, delivery requests — and a second authoring environment, which is the one thing part 18 genuinely needs.
| Starter | Pro | |
|---|---|---|
| Experience nodes | 10 | 200 |
| Components | 50 | 200 |
| Sites | 1 | 3 |
| Authoring / delivery environments | 1 / 1 | 3 / 5 |
| Categories | 100 | 5,000 |
| Blob storage | 1.5 GB | 15 GB |
| Delivery requests / day | 6,000 | 60,000 |
The series assumes the 14-day Pro trial a new organization gets, which is comfortably enough for all twenty parts. If you are reading on Starter, everything up to here fits, and everything after this behaves the same way on a smaller site.
That is the last you will hear about it until part 18.
Where this leaves us
A real catalogue, and no way to browse it. /coffees is still the empty structural node from part 4, returning a 404 to anyone who tries it.
Next: a page whose content is a query.
