---
title: "02 Exporting a Workspace"
canonical_url: https://mdflow.cz/share/c/h0L7MvX5Cy5sST3qPnApqHOmrQ36JTVy169qJMtDMya9Kc30zm04cg9OuPSSPwvj/7666542c-b167-4475-9026-2fb85f335e19
md_url: https://mdflow.cz/share/c/h0L7MvX5Cy5sST3qPnApqHOmrQ36JTVy169qJMtDMya9Kc30zm04cg9OuPSSPwvj/7666542c-b167-4475-9026-2fb85f335e19.md
visibility: public
---

# Exporting a Workspace

Download an entire workspace as a single `.zip` archive — every folder, every document's markdown, and every referenced image.

## What you get

- **The folder tree, mirrored as directories.**
- **Each document written as its unmodified markdown source.** No conversion, no wrapper, no export format.
- **Every referenced uploaded image**, bundled into one shared `assets/` folder at the archive root, with the image links in the markdown **rewritten to relative paths**.

The result is **self-contained**. Unzip it and it renders offline in any markdown tool — Obsidian, VS Code — or when viewed on GitHub. The images work because the links point at `assets/`, not at MDflow.

Each image is downloaded **once** and shared across every document that references it, so an image used in twenty documents does not appear twenty times.

## Running an export

1. Open the **workspace switcher** and start an export from the workspace you want.
2. A **confirmation step** shows how many documents and folders the archive will include.
3. **Progress is shown** while the archive is assembled, and you can **cancel at any time**.

## It happens in your browser

The whole archive is built **in your browser**, reusing the same authenticated reads as the editor. There is no separate export server, and no archive is stored anywhere.

Two practical consequences:

- **Nothing is queued and emailed to you.** The download is immediate.
- **A large workspace does real work on your machine.** Keep the tab open until it finishes.

## What is included and what is not

**Included:** folders, markdown documents, referenced uploaded images.

**Not included:**

- **Empty folders.** Folders whose entire subtree contains no documents are omitted.
- **Version history.**
- **Comments.**
- **Sharing settings**, collections, or favorites.
- **GitHub and Google Drive files** — those live in their own services and were never MDflow documents.

**Encrypted documents export as ciphertext.** MDflow cannot decrypt them for the archive, because it cannot decrypt them at all. If you want readable text, unlock and remove encryption first — or keep the archive and decrypt later with the `mdflow-crypt` CLI, which reads the same `mdflow-enc:v1` format. See [06 mdflow-crypt CLI](/doc/805092df-0030-4567-ab01-3a4c7c06aff8).

## Broken images

An image that can no longer be loaded is **skipped and noted in the summary**, and its original link is left in the document rather than silently rewritten to a file that does not exist.

## Export per workspace

Export is **per workspace**, not per account. If you have several workspaces, export each one separately.

## Why this matters

Export is the answer to the reasonable question *"what happens to my writing if MDflow disappears?"*

The answer is: you already have it. It is plain markdown in ordinary folders with working image links. There is no proprietary format to reverse-engineer, no API to scrape, and no lock-in to escape. That property is why the export exists in this shape.