# Import and export

You do not have to start a docs site from an empty page. Import a `.zip` of your existing docs, start from a template, or connect a GitHub repository. You can also export any site as plain Markdown files to back it up or move it.

## Import a .zip

### Supported sources

| Choose | What to zip |
| --- | --- |
| **Mintlify** | The folder that holds `mint.json` or `docs.json`, with its `.mdx` pages and images. |
| **GitBook** | The repository or space export that holds `SUMMARY.md`, including the `.gitbook` folder for images. |
| **ReadMe** | The folder from ReadMe's Markdown export or the `rdme` CLI, with one folder per category. |
| **Docusaurus** | The project, or just its docs folder, with `sidebars.js` and any `_category_.json` files. |
| **Markdown folder** | Any folder of `.md` or `.mdx` files. Folders become sections; images they link to come along. |
| **EvoHub export** | A `.zip` downloaded from another EvoHub docs site. It has an `evohub.json` at the top. |
| **Detect automatically** | EvoHub looks for `mint.json`, `SUMMARY.md`, `sidebars.js`, `evohub.json` and the like, and falls back to a plain Markdown folder. |

Components from those tools (callouts, tabs, cards and so on) are converted to their EvoHub equivalents where one exists. A Markdown folder follows the same layout rules as GitHub sync; see [Docs as code with GitHub](https://docs-dev.evohub.io/github-sync.md#the-repository-layout).

### Run the import

You need `docs:page:write`. Creating a new site also needs `docs:site:write`.

:::steps
### Choose the archive
In Docs, choose **Import** on the **Docs sites** page (or **Import** in an open site's menu). Drop a `.zip` of up to 50 MB, pick **Where is it from?**, and choose **Analyze archive**.

### Check what was found
EvoHub reads the archive without creating anything and shows the pages and sections it found, the number of images and files and snippets, and any warnings per file.

### Choose where it goes
- **New site** — give it a **Name** and **Address**. The pages are published right away, but the site stays private until you publish it.
- **Existing site** (or **Into this site**) — pick the **Site** and, if it has several, the **Version**. Everything arrives as drafts, in new pages and sections. Existing pages are never touched and nothing goes live. A page whose address is already taken gets a numbered address (`-2`), and links to it are rewritten. If the site requires review, one change request is opened with the imported pages.

### Import
Choose **Create site and import** or **Import into site**. A large archive can take a minute; the page updates by itself. When it is done, **Open site** (or **Open change request**).
:::

Images and files the pages link to are uploaded to the site, counted against its file storage, and identical files are stored once. Only the person who started an import can see it.

## Start from a template

On the **Docs sites** page choose **From template**, pick one, and give the new site a name and address. The template's pages (and API reference, if it has one) are published, and the site stays private until you publish it.

| Template | What you get |
| --- | --- |
| **Docs as code starter** | The layout EvoHub reads from GitHub: folders, front matter, every component, a snippet and an OpenAPI file. |
| **API documentation** | A getting-started path, guides on pagination, errors, rate limits and webhooks, and a full API reference from an OpenAPI document. |
| **Product guide** | A user guide for a product: first steps, core concepts, everyday tasks, integrations, roles and billing. |
| **Help center** | Short answers to common customer questions, behind a landing page with a search-first hero and topic cards. |
| **Developer platform** | Guides and concepts for developers, an API reference tab and a changelog linked from the header. |

Replace the example content with your own.

## Export a site

In the site's **Settings**, **General**, choose **Download as Markdown (.zip)**. The archive contains, for the version you are working on:

- every page, as the draft you are editing, with its settings as front matter;
- snippets and API references;
- the site's layout (tabs, header, footer, home page);
- uploaded images and files;
- translations, if the site has other languages;
- an `evohub.json` describing the structure.

Import the `.zip` into another EvoHub site with **EvoHub export** to reproduce the tree, titles, addresses, order, text and settings, or keep it as a backup. Anyone who can read the site can export it.

## Related

- [Docs overview](https://docs-dev.evohub.io/docs-overview.md)
- [Docs as code with GitHub](https://docs-dev.evohub.io/github-sync.md)
- [Review and change requests](https://docs-dev.evohub.io/review-and-change-requests.md)