Import and export

Imports

The administrator import workspace accepts an explicitly selected source format. The import request limit is 100 MiB. Supported file contents also share a 100 MiB uncompressed budget across all files and ZIP archives in the request.

Markdown accepts .md, .markdown, or ZIP archives containing those files. Each document requires a level-one heading (# Title), which becomes the page title; its archive/file path becomes the page path.

Wiki.js accepts JSON or ZIP. JSON pages provide path, title, and content fields.

Confluence accepts HTML/HTM or ZIP and converts a supported HTML subset to Markdown before import.

Imported pages are created as verified. When an import replaces an existing path, Lore retains the existing page’s metadata such as icon, language, tags, lifecycle state, owner/review settings, groups, deprecated target, and properties.

Exports

A single page can be exported as Markdown. If it references stored images, Lore creates a ZIP containing the Markdown plus each referenced image and rewrites image paths to archive-relative locations.

Administrators can export selected pages or all pages as an archive. PDF export uses the normal Markdown renderer, inlines stored images, and sends the self-contained HTML to the configured HTML-to-PDF service.

Filesystem static site generation is a separate publishing path described in Static sites.

Customized print and PDF output

Readers can expand Customize variables in a page’s Share and export dialog to supply temporary values for that output. Preview print / PDF and Print use the same resolved, image-embedded HTML and stylesheet as PDF export. PDF conversion still happens in the configured HTML-to-PDF service, which receives ordinary HTML and needs no variable-specific changes.

Overrides are request-local and discarded when the dialog closes. They never change stored values, pages, revisions, or Markdown exports. See Templates and snippets for the reading panel, reset behavior, and limits.

The browser-authenticated export endpoints accept JSON with the exact stored variable names. Only changed values need to be supplied:

{ "variables": { "environment": "staging" } }

POST /export/preview/{slug...} returns {"document":"..."} containing the clean print document. POST /export/pdf/{slug...} returns the generated PDF. Both use private, non-cacheable responses. Malformed requests return 400; invalid or unused variable overrides return 422 through Lore’s field-problem format. Existing GET /export/pdf/{slug...} links continue to use saved values; URL parameters do not set overrides. These are browser export routes, not bearer-token API routes.