Chapter 7 of 12

Importing & Exporting

A bibliography manager is only as useful as the references you can get into it and the formats you can get back out. Bibliofile has grown a full set of get-references-in and get-references-out paths that sit alongside the in-app Online search: you can paste BibTeX straight from the clipboard, drag and drop files onto the publications table, import from a file in BibTeX, RIS, or EndNote format, export your whole library to BibTeX, RIS, CSV, a styled HTML bibliography, or an RTF reference list — and clone the whole library, attachments and all, into a copy you can experiment on.

This chapter covers every one of those paths in depth: exactly what each one accepts, how cite-key collisions are handled, what fields are mapped where, and what lands on disk. Because importing adds entries to whichever library you have open — and, like every other edit, does not touch your .bib file until you Save — it is worth knowing precisely what each method does before you reach for it.

Note: All of the import methods in this chapter add entries to the library you currently have open. None of them replaces, merges-over, or deletes existing entries, and nothing is written to your .bib until you press Save (⌘S / Ctrl+S). If you import a batch and then quit without saving, those entries are discarded — there is no autosave by default. (You can turn on an autosave option; see Editing entries.)

7.1 The four ways in, at a glance#

Method How you trigger it Accepts Best for
Paste BibTeX Edit → Paste Publication (⇧⌘V / Shift+Ctrl+V) or Publication → New Publications from Clipboard (⌥⌘L / Alt+Ctrl+L) — the same command under two names — or just paste (⌘V) BibTeX text when the table is focused One or more @type{…} entries on the clipboard Grabbing a citation from Google Scholar, a publisher page, or an email
Drag and drop Drag files from your file manager onto the publications table .bib, .ris, EndNote .enw/.enl/.xml, PDFs, and any other file Dropping a folder of PDFs, or merging a colleague's .bib
Import from file File → Import → From File (BibTeX / RIS / EndNote)… (⇧⌘I / Shift+Ctrl+I) .bib, .ris, and EndNote .enw/.enl/.xml files chosen in a dialog A clean, dialog-driven import of one or several files
Online search File → Import → Search Online…, or the 🌐 Online… toolbar button Results from any of the six built-in sources (CrossRef, OpenAlex, PubMed, arXiv, DOI lookup, ISBN) Finding a reference you don't have a file for. See Online search.

The first three are covered here; online search has its own chapter. Whichever method you use, the same two rules apply: cite keys are kept unique (a clash gets a -1, -2, … suffix), and the import is in memory until you Save.

Every import also ends the same way, so you can see what arrived: the sidebar selection is cleared back to the whole Library (a new entry that didn't match your current group would otherwise be invisible), the table reloads, and the first newly-added entry is selected and shown in the detail pane. Any warnings from the import appear in the status bar.

7.2 Pasting BibTeX from the clipboard#

The fastest way to capture a single reference is to paste it. Most databases and search engines — Google Scholar, publisher pages, the ACM and IEEE libraries, arXiv — offer a "BibTeX" or "Export citation" button that puts an @article{…} record on your clipboard. Bibliofile turns that clipboard text into a real entry.

There are two ways to do it — a menu command and a bare paste — and both end up in the same place.

7.2.1 Edit → Paste Publication (the explicit way)#

  1. Copy a BibTeX entry to your clipboard from wherever you found it.
  2. With a library open, choose Edit → Paste Publication (⇧⌘V / Shift+Ctrl+V).

The app reads the clipboard, parses it as BibTeX, and adds every entry it finds, then selects the first of them so you can review and tidy it in the detail pane.

The same command also sits in the Publication menu as New Publications from Clipboard (⌥⌘L / Alt+Ctrl+L), where BibDesk users will expect to find it — right after the two New Publication items, since that is what it does: it makes new publications, just from text you already have. The two menu items are the same command with the same behaviour and either shortcut works; use whichever you remember. (Unlike BibDesk's version there is no ellipsis and no parsing sheet — it reads BibTeX from the clipboard and adds the entries immediately, with no further input from you.)

If the clipboard is empty — or the app can't read it at all — you get "The clipboard has no text to paste." in the status bar rather than a menu item that appears to do nothing.

7.2.2 Just paste (the quick way)#

If the publications table has focus — that is, you are not currently typing in a field, a search box, or a text area — you can simply press ⌘V / Ctrl+V. When the clipboard text looks like BibTeX (it contains something of the form @type{), the app intercepts the paste and imports it instead of doing an ordinary text paste.

Note: This "bare paste" deliberately stays out of your way while you are editing. If your cursor is in a field editor, the search box, or any text area, ⌘V pastes text into that box as normal — only a paste landing on the table (or window chrome) is treated as a BibTeX import. When in doubt, use the menu command (Edit → Paste Publication or Publication → New Publications from Clipboard), which always imports regardless of focus.

7.2.3 What paste accepts and how keys are assigned#

  • One or many entries. The clipboard may hold a single @article{…} or a whole block of several entries; all of them are imported in one go.
  • Missing cite keys are generated. An entry with no cite key is given one, derived from its author and year using your cite-key format (the same generator as the Generate button — see Editing entries). If nothing usable can be derived, it falls back to a provisional imported key you should rename.
  • Clashing cite keys are disambiguated. If a pasted entry's key already exists in your library, the app appends -1, -2, … until it is unique, so pasting never silently overwrites an existing entry.
  • Attachment blobs survive. If the pasted BibTeX already contains Bdsk-File-N attachment fields, they are carried across intact.

If the clipboard text contains no recognisable entries, the app does nothing and records the warning "No BibTeX entries found in the pasted text." — a sign the text wasn't BibTeX (a common mistake is copying a formatted citation rather than its BibTeX source).

Tip: Treat a pasted entry as a starting point. Online and clipboard BibTeX is often imperfect — ALL-CAPS titles, abbreviated venues, initials-only names. Because the first new entry is selected the moment you paste, the detail-pane editor is right there for cleanup. See Editing entries.

7.3 Drag and drop#

You can drop files straight onto the publications table — the middle column of the window, which is the import drop target. As soon as you drag files over it, an overlay covers that column reading "Drop .bib or files to import", confirming the drop will land. Release the files and the app imports them according to their type.

The drop target is that column deliberately, not the whole window, because the other panes take drops of their own: dropping a file on the detail pane attaches it to the entry you are looking at (and dropping an image on its journal cover sets the cover), while the groups sidebar takes entries dragged onto a static group. Aiming at the table is what tells the app you mean import. The overlay also appears only for files dragged in from outside — dragging a row out of the table doesn't raise it. With no library open, the welcome screen accepts a dropped .bib too, but there it opens that file rather than merging it.

You drop… What happens
.bib file(s) Each file's entries are merged into the open library, exactly as if pasted — with the same unique-cite-key handling. Warnings are prefixed with the file's name.
.ris file(s) Imported as RIS records (see §7.5).
EndNote file(s) (.enw, .enl, .xml) Imported as EndNote records (see §7.5.3).
PDF The PDF is sniffed for a DOI (then an arXiv id) and, if one is found, the reference is looked up online (CrossRef / arXiv) so a fully-populated entry is created with the PDF attached. If that paper is already in your library, the PDF is attached to the existing entry instead of creating a duplicate. With no identifier — or if the lookup finds nothing — the PDF opens the review dialog (see below) rather than being auto-added.
Any other file A new entry is created and the file attached: type = your configured default (article unless you changed it), Title = the filename (without the extension), with a generated cite key.
Plain text that looks like BibTeX Imported as if pasted (see §7.2).

This makes drag-and-drop the fastest way to build a library: drop a pile of PDFs at once and each identified paper is looked up and turned into a real entry with the PDF attached. While the lookups run, the status bar reads "Looking up references…"; when it finishes it shows a summary such as "Imported: 3 from identifier, 1 linked to existing, 2 to review."

Reviewing PDFs with no identifier#

A PDF with no DOI or arXiv id (e.g. an older scan, or a working paper that prints no identifier) isn't auto-added — that would litter a large library with empty, hard-to-find placeholders. Instead all such PDFs from the drop open a review dialog: the PDFs are listed on the left, with the normal entry editor on the right. Click a PDF to edit its entry (type, cite key, fields, notes — the title is pre-filled from the filename), then:

  • Accept — creates the entry in your library and attaches that PDF (AutoFiling it only if you have AutoFile attachments when added switched on — see AutoFile).
  • Discard — drops the PDF; no entry is created.

Nothing is added to your library until you Accept — the drafts live in a scratch area, so closing the dialog (or discarding) leaves no trace. Tip: fill in the Author, Title, and Year, then click Generate for a proper cite key.

Note: Identifiers are read from the PDF's first pages — a DOI printed in the text, or the arXiv:… watermark / abs URL on a preprint. A scanned PDF with no text layer, or a paper that prints no identifier, won't be recognized and goes to the review dialog. When you Accept, the attachment is recorded as a link (a Bdsk-File-N blob with a path relative to the library), exactly as if you had added it through the Attachments section — see Attachments.

Warning: A dropped file is attached by link, not copied into the library folder — it stays exactly where you dropped it from. For the link to keep resolving, keep the file in a predictable place relative to your .bib — or use AutoFile afterwards to move it into your Papers folder. (The one exception: with AutoFile attachments when added switched on and a Papers folder set, a dropped file is moved there immediately, as part of the same import.) See Attachments → portability.

7.4 Importing from a file#

When you want a tidy, dialog-driven import — choosing exactly which files to bring in — use the Import menu.

  1. With a library open, choose File → Import → From File (BibTeX / RIS / EndNote)… (⇧⌘I / Shift+Ctrl+I).
  2. In the file dialog (titled Import), select one or more files. The dialog filters to Bibliographies (.bib, .ris, .enw, .enl, .xml) by default; switch the filter to All Files if you need to.
  3. The chosen files are imported by type — .bib files merge their entries, .ris files import their records, EndNote .enw/.enl/.xml files import their records, and any other file you force through the All Files filter is handled just as with drag-and-drop: a PDF is looked up by its DOI/arXiv id (no-identifier PDFs open the review dialog), and any other file becomes a stub with that file attached.

The same File → Import submenu also contains Search Online (CrossRef / arXiv)… (⇧⌘O / Shift+Ctrl+O), which opens the online-search window — the fourth way in, documented in its own chapter.

Tip: Multi-selection works in the dialog (Shift-click or Cmd/Ctrl-click), so you can import several .bib/.ris files in a single operation. Each file's warnings are reported with that file's name, so you can tell which source a problem came from.

7.4.1 What survives a round trip#

A .bib you merge in — or open outright — is re-serialized when you next save, so it is worth knowing what that does and doesn't preserve. The layout is rewritten into the app's canonical shape (see Normalisations applied on save), but the content is meant to come back out exactly as it went in, including these cases that a careless parser gets wrong:

  • % comments inside an entry are treated as comments, so a % line containing a stray ,, {, or " cannot knock the field splitting out of step and drop real fields. A free-text @comment{…} or @preamble{…} body is left alone instead, so a literal % in one (JabRef and Zotero emit these) survives untouched.
  • A missing } damages only the entry it belongs to: the parser resynchronises at the next @type{ that starts a line, rather than swallowing every entry after it.
  • Unmatched braces in a value — a title containing a lone { or } — are written out as the balanced TeX control words {\textbraceleft} and {\textbraceright}, which re-read as the same characters and are shown as { and } in the app. Values whose braces are already balanced (which is nearly all of them) are written through unchanged.

7.5 The RIS format#

RIS is a long-standing tagged citation format exported by EndNote, Zotero, Mendeley, PubMed, ScienceDirect, and many library catalogues. A RIS record is a list of two-letter tags, one per line, bracketed by a TY (type) line and an ER (end of record) line:

TY  - JOUR
AU  - Einstein, A.
AU  - Podolsky, B.
AU  - Rosen, N.
TI  - Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?
JO  - Physical Review
PY  - 1935
VL  - 47
SP  - 777
EP  - 780
DO  - 10.1103/PhysRev.47.777
ER  -

When you import RIS (by drag-and-drop, by File → Import, or by dropping a .ris file), each record becomes a BibTeX entry. A trailing ER is optional — the last record is imported even if its ER line is missing. As with every other import, cite keys are generated and de-duplicated automatically, and the library is marked unsaved.

7.5.1 RIS type → BibTeX type#

The RIS reference type (TY) chooses the BibTeX entry type:

RIS TY BibTeX type
JOUR article
BOOK book
CHAP incollection
CPAPER, CONF inproceedings
THES phdthesis
RPRT techreport
UNPB unpublished
GEN misc

Any reference type not in the table is imported as misc. (You can correct the type afterwards from the Type dropdown — see Editing entries.)

7.5.2 RIS tag → BibTeX field#

RIS tag(s) BibTeX field Notes
AU, A1 Author Multiple lines accumulate, joined with and
A2, ED Editor Multiple lines, joined with and
TI, T1 Title
T2, JO, JF Journal or Booktitle Journal for a JOUR record, Booktitle otherwise
PY, Y1 Year A leading four-digit year is kept and the rest dropped (1935/01/01 → 1935); a value that doesn't start with four digits is taken as-is
VL Volume
IS Number
SP / EP Pages Combined as start--end; a lone SP becomes a single page
PB Publisher
CY Address
SN Isbn
DO Doi
UR Url
AB, N2 Abstract
KW Keywords Multiple keywords, joined with ,

Tags the importer does not recognise are ignored, so an unusual RIS file imports cleanly with its mappable fields and quietly drops the rest.

7.5.3 EndNote import#

Bibliofile reads the two EndNote formats you actually run into:

  • Refer / tagged (.enw) — the line-oriented %X value format. This is what Google Scholar's "EndNote" download produces, and what many journal sites offer as "Download citation → EndNote". Records are separated by blank lines; %A (author), %E (editor) and %K (keywords) may repeat.
  • EndNote XML (.xml, .enl) — EndNote's richer <xml><records>… interchange export. Text wrapped in EndNote's <style> runs is unwrapped to plain text automatically.

The reference type sets the BibTeX type (Journal Article and Electronic Article → article, Book and Edited Book → book, Book Section → incollection, Conference Proceedings/Paper → inproceedings, Thesis → phdthesis, Report → techreport, Manuscript and Unpublished Work → unpublished; anything else, including Web Page and Generic, → misc). Common fields map across as you would expect:

EndNote (tagged / XML) BibTeX field Notes
%A / <author> Author Joined with and
%E / <secondary-authors> Editor Joined with and
%T / <title> Title
%J / — Journal The tagged format has separate tags, so no guessing is needed
%B / — Booktitle
— / <secondary-title>, <full-title> Journal or Booktitle XML has one venue element: Journal for an article, Booktitle otherwise
%D / <year> Year A four-digit year is picked out of the value
%V / <volume> Volume
%N / <number> Number
%P / <pages> Pages A single hyphen between numbers becomes --
%I / <publisher> Publisher
%C / <pub-location> Address
%R / <electronic-resource-num> Doi A leading doi: is stripped
%U / <related-urls> Url
%@ / <isbn> Issn or Isbn Issn when it matches the ISSN pattern, else Isbn
%K / <keyword> Keywords Joined with , ; a single %K line holding several comma- or semicolon-separated keywords is split into separate keywords first
%X / <abstract> Abstract
%Z, %1 / — Note
%G / — Language

As with RIS, EndNote imports get fresh generated cite keys and are added as new entries (nothing is overwritten); unrecognised tags are ignored.

7.6 Exporting your library#

To get references out of Bibliofile in a different format, use the Export menu. Each format writes the whole library to a file you choose. (This section also covers the two neighbouring ways out that aren't file formats at all: printing and cloning the whole library.)

Choose File → Export → and then one of:

Format Menu item What it produces
BibTeX BibTeX… A standard .bib file in BibDesk's canonical format — the same serialization a normal Save produces, including macros and the document header.
RIS RIS… A .ris file with one record per entry (see the mapping below).
CSV CSV… A comma-separated spreadsheet with a fixed set of bibliographic columns.
HTML HTML… A self-contained, styled HTML bibliography page (rendered through a Handlebars template).
RTF RTF (formatted bibliography)… An .rtf document you can open in a word processor: one CSL-formatted reference per paragraph, in your default citation style, with the italics that style calls for. See §7.6.5.

Each choice opens a Save dialog (titled Export) with a sensible default filename — your library's name with the appropriate extension. None of the export items has a keyboard shortcut; they live entirely in the File → Export submenu.

To export just some entries, select them in the table and choose File → Export → Selected Entries (BibTeX)…. This writes only the highlighted entries (or, if exactly one row is selected, that one) to a .bib file — handy for pulling the references for a single paper out of a large library. The default filename is your library's name with a -selection suffix.

If you have defined any export templates of your own, they appear at the bottom of the same File → Export submenu — each as a submenu offering Whole Library…, Shown Entries…, or Selected Entries…. That is the route to a layout of your own design, and the only export route that can write just what the table is currently showing. See Customizing panels → Export templates.

Note: The five built-in Export formats always write the entire library, regardless of the group you have selected or any live-search filter — they are export the whole file operations. Use Selected Entries (BibTeX)… (above) when you want only the highlighted entries, an export template for other scopes, or Edit → Copy as BibTeX to put a single entry on the clipboard — see Editing entries.

Tip: Exporting to BibTeX is also a convenient way to take a one-off snapshot of the library to a new file without disturbing your working .bib or its .bib.bak backup. To save the working library under a new name and keep editing it there, use File → Save As… (⇧⌘S) instead; to get a snapshot that carries copies of the attached files as well, use Clone Bibliography….

7.6.1 RIS export#

Each entry is written as a RIS record. The BibTeX type is mapped to a RIS TY (for example article → JOUR, book → BOOK, inproceedings → CPAPER, phdthesis → THES, misc → GEN; unknown types fall back to GEN), and the fields are emitted as their RIS tags — AU per author, A2 per editor, TI, T2 (journal or book title), PY, VL, IS, SP/EP split from Pages, PB, CY (from Address), SN (from Isbn/Issn), DO, UR, AB, a KW per keyword, and ID carrying the cite key. Values are de-TeXified on the way out so accents and braces export as plain Unicode text.

7.6.2 CSV export#

The CSV has a fixed header row and one row per entry, with these columns:

Cite Key, Type, Authors, Title, Year, Journal, Volume, Pages, Publisher, DOI

The Journal column falls back to the book/proceedings title when an entry has no journal, multiple authors are joined with ; , and cells containing commas, quotes, or newlines are quoted and escaped to standard CSV rules. The result opens cleanly in any spreadsheet program.

7.6.3 HTML export#

The HTML export produces a complete, self-contained web page — a styled bibliography you can open in a browser, email, or publish. The page has its own inline styling (so it needs no external CSS), a heading showing the library's name and the number of entries, and one formatted reference paragraph per entry:

Einstein, A., Podolsky, B., Rosen, N. (1935). Can Quantum-Mechanical
Description of Physical Reality Be Considered Complete?. Physical Review,
47, 777. https://doi.org/10.1103/PhysRev.47.777 [einstein1935]

Each reference shows the authors, the year, the title, the venue, volume and pages, a clickable DOI link, and the cite key in brackets — and any field an entry lacks is simply left out of its line. The page is generated from a Handlebars template, the same lightweight templating approach BibDesk's old custom .template language served, and every interpolated value is HTML-escaped, so titles with <, >, or & export safely.

Note: The HTML bibliography is a fixed, readable reference layout — not the CSL-formatted citation you see in the detail pane, and not a bibtex/biber bibliography. It is meant for sharing a glance-able list, not for typesetting a paper. For a properly styled single citation, use the formatted-citation block or Edit → Copy Citation; for a whole bibliography in your chosen reference style, export RTF or print it instead.

7.6.4 Printing a bibliography#

File → Print… (⌘P / Ctrl+P) produces a printed, CSL-formatted bibliography — the same styled citations you see in the detail pane, one per entry, with a hanging indent and a heading. Choosing it opens your system's standard Print dialog, which on macOS also offers Save as PDF (a quick way to get a formatted PDF reading list without exporting a file first).

Unlike Export, Print follows what you are looking at:

  • If you have two or more entries selected, only those are printed.
  • Otherwise everything the table is currently showing is printed — the current group (the whole library, or a static or smart group's members) narrowed by whatever is in the live-search box.

The heading is your library's name, and — when a group other than the whole library is selected — the group name too (for example math-demo — To Read). The entries are formatted with your default citation style (set in Preferences), so switching styles there changes how the printout reads.

Note: Print uses the CSL citation formatter, not the HTML-export template — so a printout matches your chosen reference style, whereas File → Export → HTML… always writes the same fixed, glance-able layout.

7.6.5 RTF export#

File → Export → RTF (formatted bibliography)… writes the same CSL-formatted references that Print produces, but to an .rtf file instead of the printer — one reference per paragraph, formatted with your default citation style, ready to paste into a manuscript or hand to a co-author who doesn't use BibTeX. The italics and bold the style asks for are carried into the RTF; links become plain text, since RTF has no clickable-DOI equivalent here.

Unlike Print, this is one of the whole-library exports: it writes every entry, not the current selection or view.

7.6.6 Cloning a bibliography#

File → Clone Bibliography… makes a complete, self-contained copy of the open library — including its attached files — and opens it in a new window. It exists for the moment when you want to try something drastic (a big Find & Replace, a re-run of AutoFile, OCR over a pile of scans) without risking the library you actually depend on.

  1. Choose File → Clone Bibliography…. A Save dialog (titled Clone Bibliography) opens, defaulting to «your library» copy.bib beside the original.
  2. Pick a name and location. The clone is written there, and every file the library links is copied into your Papers folder, named by your AutoFile format exactly as a freshly added file would be — see AutoFile. With no Papers folder set, the copies go into a «clone name» Papers folder next to the clone instead, and the summary tells you so.
  3. The clone's attachment links are rewritten to point at those copies, and the clone opens in its own window.

A summary then reports what happened — "Cloned 412 entries to 'thesis copy.bib'" — along with how many attachments were copied and the folder they went into, and a warning list if anything could not be copied.

What the clone guarantees, and why it is not just "Save As":

  • The original is untouched. Its file, its unsaved-changes state, its undo history, and — crucially — every file it links are left exactly as they were. Your open window keeps pointing at the original, not at the clone (that is what Save As does instead).
  • No existing file is overwritten by the copies. An attachment copy always gets a fresh, non-colliding name, even when a file of that name is already sitting at the target — so a clone can never end up sharing a PDF with the original. And cloning onto a .bib that is currently open (the source itself, or another window) is refused outright. (The save dialog will still ask before replacing some other, closed .bib you point it at, as usual.)
  • Unsaved edits come along. The clone is written from what is in memory, so you needn't save the original first.
  • Files it can't copy are reported. If an attachment is missing from disk or can't be read, the clone keeps that link pointing at the original file — a live link beats a broken one — and the summary says so. Those particular files are shared with the source library, so treat them as read-only in the clone.
  • The encoding is preserved. The clone is written in the source's text encoding; if a rewritten path contains characters that encoding can't hold, the summary warns you rather than silently mangling it.

Tip: A clone is also a reasonable way to hand a library to someone else: every file it links is a copy (bar any the summary reported as uncopyable), not a file borrowed from elsewhere on your disk. Its links are stored relative to the clone, though, so it travels intact only if the clone and the folder its copies went into keep the same relative positions. With no Papers folder configured, those copies land in a folder right next to the clone — the easiest arrangement to zip up and send.

7.7 Quick reference#

Action How
Paste BibTeX (explicit) Edit → Paste Publication (⇧⌘V / Shift+Ctrl+V) or Publication → New Publications from Clipboard (⌥⌘L / Alt+Ctrl+L)
Paste BibTeX (quick) Copy BibTeX, then ⌘V / Ctrl+V with the table focused
Drag-and-drop import Drag .bib/.ris/EndNote/PDF/other files onto the publications table
Import from a file File → Import → From File (BibTeX / RIS / EndNote)… (⇧⌘I / Shift+Ctrl+I)
Search online File → Import → Search Online… (⇧⌘O) or 🌐 Online…
Export BibTeX File → Export → BibTeX…
Export RIS File → Export → RIS…
Export CSV File → Export → CSV…
Export HTML File → Export → HTML…
Export RTF File → Export → RTF (formatted bibliography)…
Export selected entries File → Export → Selected Entries (BibTeX)…
Export through your own template File → Export → «your template» → Whole Library… / Shown Entries… / Selected Entries…
Clone the library + its files File → Clone Bibliography…
Print / Save as PDF File → Print… (⌘P / Ctrl+P) — prints the selection, or everything the table is showing
Persist imports Save (⌘S / Ctrl+S) — imports are in memory until saved

7.8 Troubleshooting#

"Paste Publication did nothing / 'No BibTeX entries found.'" The clipboard didn't contain BibTeX. Make sure you copied the BibTeX source (@article{…}), not a formatted citation or an HTML snippet. Re-copy from the source's "BibTeX" / "Export citation → BibTeX" option and paste again.

"A plain ⌘V didn't import — it pasted text instead." The bare-paste import only fires when the table (not a text field) has focus and the clipboard looks like BibTeX. If your cursor was in a field, the search box, or a text area, ⌘V pastes there as normal. Use Edit → Paste Publication (or Publication → New Publications from Clipboard, ⌥⌘L), which imports regardless of focus.

"I dropped files on the window and nothing happened." Only the publications table — the middle column — is the import drop target. A file dropped on the detail pane is attached to the entry you are looking at instead, and one dropped on the groups sidebar does nothing. Watch for the "Drop .bib or files to import" overlay: if it isn't showing, the drop won't import.

"I dropped a PDF but the entry has no real title/author." If the PDF prints a DOI or arXiv id in its first pages, the entry is built from a real online lookup. If it doesn't (e.g. an older scan), the PDF opens the review dialog instead of being auto-added — fill in the fields there before you Accept, or look the paper up with Online search and copy the fields across. (Anything you Accept with only the filename title can be edited later, too.)

"My import isn't in the file after I quit." Imports are unsaved edits. Press ⌘S / Ctrl+S (or click Save) after a batch of imports. There is no autosave unless you enable it in Preferences.

"The RIS import dropped some fields." Only the RIS tags in the mapping table are imported; unrecognised tags are ignored. Add any missing fields by hand in the detail pane.

"My export looks reformatted / lost my Markdown." The CSV, RIS, HTML, and RTF exports de-TeXify and flatten values for their target format, so braces and TeX accents become plain text and Markdown in abstracts is not rendered. This is expected — those formats can't carry BibTeX/Markdown markup. Your .bib library itself is untouched by exporting.

"Clone Bibliography wouldn't write where I asked it to." A clone is refused if the file you picked is a bibliography that is currently open — including the original itself. Close that window (or pick another name) and try again. The refusal is deliberate: the open library's in-memory state would overwrite the freshly written clone the next time it saved.

"The clone says some attachments couldn't be copied." Those files were missing from disk or unreadable, so there is no copy to link. The clone keeps those particular links pointing at the original files, which means editing or re-filing them from the clone will affect your real library — the summary lists exactly which entries are affected. Fix the links (see Find Broken Links) and clone again if you want a genuinely independent copy.

See also#

  • Online search — the fourth way in: pull references from CrossRef, OpenAlex, PubMed, arXiv, or a DOI/ISBN lookup.
  • Editing entries — clean up an imported entry, and the Copy as BibTeX / Copy Citation clipboard commands.
  • Attachments — how dropped PDFs are stored, and AutoFile (which decides where a clone's copied files land).
  • Customizing panels — writing your own Handlebars export templates, which then appear in File → Export.
  • Scripting with JavaScript — import, export, and transform entries programmatically (doc.import, doc.export, doc.save) when the menu commands aren't enough.
  • Shortcuts & reference — the full menu and keyboard reference.