Import

File > Import brings an existing manuscript into the currently open project as binder documents/folders, landing under whichever folder is selected in the binder (or the project root if nothing is). Four formats are supported, each with a different picker (a single file for the first three, a whole folder for Scrivener) and a different level of fidelity, since the formats themselves differ in how much structure there is to recover:

• Word Document (.docx) — split into one document per “Heading 1”-styled paragraph (the standard Word convention for chapter titles), falling back to a single document if the file has none. Bold/italic/strikethrough survive; a table is flattened to one plain paragraph per cell rather than reconstructed as a markdown table.
• EPUB — one document per chapter, in the book’s own reading order (its spine, a well-defined boundary unlike DOCX’s heading heuristic), titled from each chapter’s own first heading. Bold/italic/strikethrough/code survive, along with headings, blockquotes, and lists.
• Scrivener Project — pick the .scriv project folder itself (it’s a directory, not a single file). Its Draft/manuscript folder maps onto smaragd’s own Manuscript role; Trash is skipped entirely rather than imported. Bold/italic survive. This is the newest and least-tested of the four importers — if something looks wrong after importing a real Scrivener project, that’s expected territory, not user error.
• PDF — always a single document, plain text only: no formatting, no chapter splitting. A PDF has no semantic markup to recover in the general case (just glyphs positioned on a page), so this is inherently the lowest-fidelity option of the four — reach for it only when nothing better is available.