You can download a copy of your published site and your collection data at any time. The export is a ZIP file that works on any static host. It is not a live copy of Remake, and it does not include the editor.
Last updated
Only a project owner or admin can download the export, and the project needs a hosting plan that includes site export (Starter and above). On the Launch plan you can build and preview, but not download.
The ZIP contains:
Every page in the export source as HTML: the published version, or the current draft when the project has never been published.
The CSS for each page, included in the page.
The images those pages use, in a media/ folder.
One CSV file for each collection, in data/.
One template for each collection list page and item page.
One JSON translation file for each additional language, in translations/.
Your favicon and robots.txt, when the project has them.
A README.md explaining the format and the limits.
Open index.html to view the home page. You can upload the whole folder to a static host, or keep it as an archive. Most hosts use 404.html automatically; if yours does not, set it as the not-found page.
Open the project, go to Settings → Project settings → Export site, and choose Download export.
The export is built from the latest published version. Draft changes that were never published are not included.
If the project has never been published, the export uses your current draft instead. Draft exports are clearly marked with a banner and a draft filename. They are useful for review, backup, or moving work in progress, but they are not the version visitors see.
You can download one export per project every 24 hours. The page shows the last export and when the next one is available. The limit is per project, so asking a different team member does not reset it.
A collection can have thousands of items. Exporting one HTML file for every item would create a huge archive, and many static hosts limit the number of files. Remake does not do that.
Instead:
The collection list page is exported with one sample item.
The collection item page is exported once as _item.html.
The CSV contains every item in the collection.
The sample is wrapped in HTML comments, so it is easy to find and replace:
<!-- remake:collection=blog-posts -->
<!-- remake:repeat=item -->
<article>
...
</article>
<!-- /remake:repeat -->
On the item template:
<!-- remake:item=blog-posts -->
<h1 data-field="Title">Spring collection</h1>
<img data-field="Cover image" src="media/spring-cover.webp" alt="">
<!-- /remake:item -->
data-field uses the label you gave the field in Remake. It never uses an internal field id.
To make the dynamic pages live elsewhere, replace the marked sample block with a loop over the CSV or your own database. The sample shows the real markup, classes, and field names.
If a collection has no published items, its templates use clearly labelled sample values so the page is not empty. The README says those values are not real content.
Each collection gets its own CSV file. Column names are the labels you defined, not internal ids.
The first columns are:
Name: the item name.
Slug: the item slug.
Status: Published, Scheduled, or Draft.
Published: the publishing date.
Then come your own fields, in the order they appear in the collection, followed by Created at and Updated at.
Values are written in a readable form:
Select fields export the option labels, not option ids.
Reference fields export the linked item names. If two items share a name, the slug is added in parentheses, for example Ada Lovelace (ada).
Image, file, and video fields export the original file name.
Rich text exports as HTML inside the cell.
Dates export in ISO format.
Switches export as Yes or No.
The CSV includes drafts and scheduled items. That is deliberate: it is a copy of the collection, not just the live part.
The export is static. Anything that needs Remake to be running will not work:
Forms do not submit.
Workflows and scheduled jobs do not run.
Search, filters, and repeater pagination do not query live data.
End-user sign-in, protected pages, account pages, and password reset do not work.
Comments, the media library, and the editor are not included.
Changes you make in Remake after the export do not appear in the files.
The export leaves out anything private or account-related:
Passwords, sessions, OAuth secrets, and other credentials.
End-user accounts and profile data.
Billing, subscriptions, invoices, and payment data.
Analytics events and visitor data.
Comments and AI conversation history.
Deployment history and workflow definitions.
Domain and DNS settings. Your domain stays with your registrar.
The HTML pages are in the project's default language. Other languages are exported as JSON files in translations/, not as a second set of HTML pages. This keeps the archive small and gives you the strings and content you need to rebuild the localized pages.
Each translation file contains:
messages: the source text and its translation, with a reviewRequired flag when the source changed after the translation was written.
pages: translated page titles and descriptions, keyed by page path.
collections: translated select option labels and published item translations, keyed by collection slug, item slug, and field label.
siteMessages: translated visitor-facing system messages, when the project has any.
Inline variables keep their position as {{variable}} markers. Only published item translations are included. Unpublished translations stay in Remake.
Uploaded images referenced by the exported pages and collection fields are downloaded into media/. External images that are not hosted by Remake stay as external URLs.
A project with a very large media library produces a larger ZIP and can take longer to prepare. Exports are limited to 256 MB. If an archive is larger, the download is refused with a message instead of producing a partial site. Remove unused media or contact support for a manual export.
The CSV still contains every item; the archive never silently drops rows.
Internal links are relative, so the folder can be moved or uploaded without changing a domain name.
A link to a page that was not exported, such as a protected page, may point to a missing file. The README lists pages that were skipped and why.
The export gives you the published result and your content. It does not contain the Remake editor format, page documents, collection settings, or a file you can import back into Remake. If you need a copy that can be restored into Remake, contact support instead of relying on the export.
A website builder that works like a document. CMS, forms and EU hosting included.
© 2026 Remake. All rights reserved.
