Guide
How to recover a deleted Framer CMS item
If you have deleted a CMS item in Framer and need it back, this guide explains exactly what your options are — and, honestly, what your options are not.
Before anything else: don't publish yet
Deleting an item in the editor does not change your live site. Framer's help centre is explicit that edits only become visible after someone publishes. So if nobody has published since the deletion, the item is most likely still on your live site.
- Ask everyone with publishing rights to hold off, and pause any scheduled-publish plugin or automation until you have what you need.
- Open the item's page on your live domain (not staging, if you published there separately). Save it as a web page or PDF and copy the text into a document. After pasting it back into a rich text field, check headings, links and lists.
- Note the slug, SEO title and description, and the image addresses (right-click, copy image address) so you can rebuild the item with the same URL.
- If the collection has no detail pages, look for the content where it appears on your live site, such as a list, a grid or a testimonials section.
Recreate the item in the editor, check it on staging or in preview, and only then publish. A live page shows what visitors see, not every field: drafts, hidden fields and reference IDs will not be there.
Was it deleted on a branch, or in an earlier publish?
Since Framer 3.0, branches carry CMS items and collections, and changes made by Framer's Agent or an external agent (Claude Code, Codex, Cursor) happen on a branch. If the deletion is on a branch that hasn't been merged, the item is still on Main: don't merge that branch until it's fixed. More in our guide to Agents, Branching and the CMS.
If the deletion was already published, look under Staging & Versions in your site settings. Framer says versioned links there are static and unaffected by future publishes, so a version published before the deletion may still show the item's page. Copy the content from it the same way as from a live page.
What does Framer's version history actually save?
Framer's version-history guide explains how to copy earlier canvas elements into the current project. It does not document a deleted-CMS-item recovery workflow. Do not assume that recovering a layout also recovers its collection data. Check the collection itself before publishing.
If the deletion just happened, pause other edits and check whether Undo is still available in the editor. Preserve any existing exports and the current content before trying a recovery. Avoid repeated bulk imports while investigating.
The Framer Server API and what it means for recovery
Framer exposes a Server API (currently in beta) that lets you read and write CMS collections programmatically. The API provides methods to fetch all items in a collection, add new items, and remove items.
There is one critical constraint: deleted node IDs cannot be resurrected. When an item is deleted in Framer, its internal node ID is gone. If you try to create a new item with the same node ID via the API, Framer will create a fresh item with a new ID instead. This means that even if you have a complete copy of the item's data, the restored item will have a new ID — and references should be checked after recovery. FrameVault remaps references where it can resolve them; unresolved links may need manual attention.
Your recovery options, in order of practicality
Option 1: A dedicated backup tool (most reliable)
If you were using an automated backup solution before the deletion, recovering the item is straightforward: find the snapshot that contains the item, view its field values, and use the restore feature to write it back to Framer via the Server API.
The restore call uses the Server API to recreate saved field values. Image and file fields contain URLs, not separately archived media files. Because of the node ID constraint described above, Framer will assign a new ID. FrameVault reports whether an item was recreated. Check related items afterward; a new-ID notice is not a complete audit of incoming references.
Without an existing backup, this option is not available to you. But if you are reading this guide after recovering from a deletion, this is the main reason to set up backups now.
Option 2: The Framer CMS Export plugin (manual, point-in-time only)
Framer has a first-party CMS Export plugin that lets you download your CMS data as a CSV. If you or someone on your team ran an export recently, you may have the item's data in that file.
The limitation is that this export is entirely manual — it only exists if someone explicitly ran it before the deletion. There is no automatic history. And Framer supports importing CSV files back into a CMS collection. Follow its CSV import guide, check the field mapping and try a small sample in a disposable project first. Review slugs, images, rich text and references before importing into the live collection.
To take a manual snapshot of every collection at once before your next big edit, our free FrameVault Backup plugin downloads the whole CMS as one JSON file, or one collection as CSV. It has the same limit: it only helps if someone ran it before the deletion.
Option 3: Contact Framer support
Ask Framer support whether recovery is available for your specific incident. Include the project, collection, approximate deletion time and what is missing. We cannot guarantee that support has a recoverable copy or estimate the chances of recovery. Keep any earlier exports while you wait.
Option 4: Reconstruct from memory or secondary sources
If none of the above apply, your last option is manual reconstruction: re-enter the item's data from memory, from a Google Doc, from your own content drafts, or from a previously saved page or an archived public page in the Wayback Machine. Public pages can help recover visible text; they do not contain every CMS field. This is time-consuming and often incomplete, especially for structured fields like dates, references, or rich text.
What to do right now if the deletion just happened
- Don't publish. If nobody has published since the deletion, save the item's content from your live site first (see the box at the top).
- Pause edits and check whether editor Undo is available. Then ask whether anyone on your team has a recent manual CSV export from the Framer CMS Export plugin.
- If it was already published, open an earlier version's link under Staging & Versions in site settings and copy the item from there.
- Preserve any already-open page or saved copy that still shows the content. Avoid refreshing that copy until you have saved the visible text.
- For public content, check whether the Wayback Machine has a copy of the page. An archived page may be missing media and structured fields.
- Contact Framer support and describe the deletion — include the project name, collection name, and the approximate time of deletion.
- Set up automated backups before the next deletion happens (more on this below).
After recovering: preventing the next one
The most effective prevention is an automated backup that runs without requiring anyone to remember to export. The Framer Server API makes this possible: a server-side process can read your entire CMS at regular intervals and store the result in persistent storage.
What a good automated backup gives you:
- A timestamped snapshot with collection counts and clearly stated coverage limits
- The ability to browse the history and see what the item looked like at a saved recovery point
- A tested restore path that reports reference problems and new item IDs
- CSV and JSON exports for any snapshot, any time
FrameVault: automated Framer CMS and site backups
FrameVault connects to your Framer project via the Server API and takes automatic weekly snapshots (daily on paid plans). Browse the full history, compare snapshots field by field, download JSON or CSV, or restore a deleted item directly from the dashboard. The Free plan is available without a credit card; restore requires a paid plan. CMS snapshots keep your content and media links, not separate copies of media files. Pro, Studio and Agency also save a site archive of every published page.
Note: item restore creates a new node ID due to a Framer Server API constraint (beta). FrameVault surfaces a clear warning when this applies.