Skip to content

Features

Keys are written as on Windows and Linux. On a Mac, press ⌘ where it says Ctrl and ⌥ where it says Alt. The app shows each key as your computer has it.

Writing sessions

A session lives in a group channel and holds multiple prompts. Anyone in the group can create a session. Each prompt can be started, timed (10 seconds to 24 hours), and stopped early by its creator or a group admin.

  • While a prompt runs you write on a page of paper in the type a document is written in. Your writing stays private until the prompt ends.
  • A bar stays in sight above the page as you scroll. It holds the clock, your word count, your pace in words a minute and whether your words are saved, with the tools under it: bold, italic, underline, quote, lists, divider, undo and redo. Headings, strikethrough and pictures are under More. On a phone the tools sit on the keyboard.
  • Only the prompt and the page: the arrows button at the end of the bar fills the window with the prompt and your page and nothing else. Escape goes back. The prompt can be put away there too.
  • Quotes, apostrophes, dashes and ellipses are a typographer's as you type them, as in a document. The switch for that is in Settings ▸ Writing.
  • Your words are kept safe at the bell. Writing is saved a moment after you pause, sooner in the last twenty seconds, and once more just before time is up. A save that does not get through is said in the bar and sent again. A copy is also kept on this device the whole time. If the app closes mid-sprint, the page comes back as you left it. If your last words did not reach the server before time ran out, the session says so and lets you copy them or save them as a document. The others cannot see words the server never got.
  • A line under the page says who else is writing right now.
  • When a prompt ends everyone's writing is shown, your own piece first, each on its own page with its word count. A prompt that ended a while ago stays folded so a long session is not an endless page.
  • Copy takes any piece as plain text. Save as a document turns your own piece into a Document of yours (on this device in the apps, on the server in a browser), where everything else in the app works on it.
  • Vote on the pieces. When two or more people wrote for a prompt, everyone in the session can vote for the piece they liked best. It can't be their own. A vote can be moved to another piece or taken back while the voting is open. The session shows how many have voted but never who voted for which piece. Whoever started the prompt, the session's creator or a group admin closes the voting, and then each piece shows its votes with the most voted first. Ending the session leaves the voting open.
  • The prompt editor has the full formatting toolbar (headings, lists, quotes, inline code, pictures) and a choice of timer lengths.
  • A join card is posted to the channel when a session is created, so members can jump in from chat.
  • Ended sessions stay browsable forever. The creator or an admin can also delete a session permanently.

Documents

Documents are for Google-Docs-style collaborative writing and are separate from timed sessions. They are also a book-production tool: write a manuscript, design the book, export a print-ready paperback, an ebook and a submission manuscript.

  • Live multi-user editing: a CRDT (Yjs) syncs everyone's edits and shows live cursors. There is no lock, no conflict and no "someone else is editing" wall.
  • Your text is never lost: edits that can't reach the server yet (offline, a dropped connection, quitting mid-sync) are kept on this device and sent when the server is back. They're never sent to the wrong server or the wrong document. The save indicator says Saved, Saving…, Offline or Not saved (with what to do about it), and closing, signing out or quitting waits a moment for unsent edits. On-device documents save atomically, so a crash mid-save can't damage the file.
  • Writing formats: None, Screenplay, Stage Play, Manuscript, or Poetry. Each format adds its own element types (scene heading, character, dialogue, …) with correct margins and Final-Draft-style Tab/Enter cycling between them.
  • Manuscript is the book format: elements for Chapter Heading, Part Heading, Unnumbered Chapter (Prologue, Epilogue, Acknowledgments…), Chapter Subtitle, Subheading, Epigraph, Attribution, Verse and Body (Ctrl+1 to Ctrl+9 set them). Chapters number themselves ("Chapter Three" labels show beside each heading). Scene breaks are * * * + Enter or the toolbar button. A page break (Ctrl+Enter, or the toolbar's page-break button) starts what follows on a new page (a letter, a poem, an interlude on a page of its own) in the print PDF, the ebook and the submission manuscript alike. Chapters start on a new page by themselves. You write meaning rather than layout: indents, fonts and spacing come from the book's design, so editing the text can't break the book. In Manuscript, Tab only changes an empty line's element (it never turns a sentence into a heading), and pasted headings and *** lines become chapters and scene breaks.
  • Smart quotes and dashes as you type (Settings ▸ Writing to turn off): curly quotes in the style of the text's language, -- becomes an em dash, and ... becomes an ellipsis (…). The text's language is the book's for a Manuscript. For any other document it is the language set in its Details, and this device's language when none is set.
  • AutoCorrect (Settings ▸ Writing, "Correct as I type"): when a word is finished with a space or Enter, a slip is set right (teh, recieve, dont), a sentence gets its capital, TWo capitals and a wORD typed with the caps lock on are mended, and (c), 1/2 and 1st become ©, ½ and 1st with its letters raised. One Undo, or Backspace straight after, takes back the correction and leaves what you typed. Each kind has its own switch, and Your corrections… lists every entry, lets you add your own and take any out, and holds your exceptions.
    • It is careful. A word is only changed once you have left it. A slip that is part of an address or a file's name is left. A capital in the middle of a sentence is taken for a name and left as spelled. Nothing is changed in code, and while you are suggesting only your own suggested words are touched.
    • The built-in words are English and apply to English text (with "licence" and "organisation" where the document's English is British). Signs and your own entries apply in every language.
    • A script's lines of capitals, and verse, get no capital of AutoCorrect's. On a phone, the list and the capitals start switched off, since the phone's keyboard already does both.
    • The English list comes from LibreOffice's (see the notices), with every entry left out whose "slip" is a real word.
  • Bold, italic and underline reach the whole word when the caret is inside one and nothing is selected. At a word's edge or in a space they set what you type next, as before.
  • Lists start as you type them in a Document: 1., 1), a., A., i. or I. at the start of an empty paragraph begins a numbered list of that kind, and * or - a bulleted one. Start at c. or 3) and the list starts there. A label typed under a list that it continues joins that list.
  • Spaces and hyphens that hold a line together: a no-break space (Ctrl+Shift+Space, or Option+Space on a Mac) keeps two words on one line, a no-break hyphen (Ctrl+Shift and the hyphen key) is never broken at, and an optional hyphen (Ctrl and the hyphen key) shows only where the line breaks there. All three are also in Insert ▸ Symbol, by name. Find sees through them, Word files keep them, and print and PDF honor them.
  • Move a paragraph by key: Alt+Shift+Up or Down (Ctrl+Shift with the arrows on a Mac) moves the paragraph the caret is in, every paragraph a selection touches, or a list item with all that is under it. A selected picture moves the same way. Each step is one undo. While you are suggesting, switch to Editing to move things.
  • Make it a book manuscript: switching a Document to Manuscript (or Export ▸ Make it a book manuscript…) detects chapters, parts, scene breaks and front matter and shows what it found before converting. One undo puts it back. Importing a Word manuscript does the same automatically.
  • Book inspector (the book button, Manuscript): the book's details (title, author, series, ISBNs, copyright, dedication, epigraph, also-by, about the author, submission contact) and its design. The design has five presets (Classic, Literary, Fantasy, Romance, Modern) with every choice adjustable: fonts, size, leading, chapter labels, drop caps, small-caps lead-ins, ornaments, running heads, page numbers. The body and heading fonts can be any of the app's seven book faces or any family of the online library (the same Google Fonts collection documents use): it downloads when chosen, draws the manuscript view and the preview, and is embedded in the paperback PDF. Other scripts of the family (Cyrillic, Greek…) are fetched when the text has them, and where a display face lacks an alphabet a title in it is set in the body face. A library family has no small capitals or old-style figures of its own (capitals are reduced instead), and one with no italic is slanted. If it can't be fetched (offline and never downloaded) the book is set in the nearest of the app's own and the export says so. Also here: print settings (trim size, KDP-safe margins that grow with the page count, right-hand chapter starts, front and back matter, spine width) and ebook settings. Saved in the document, so collaborators see the same book.
  • Your own scene break (Book ▸ Design ▸ Scene breaks): one of the typed ornaments, characters of your own (a dagger, a cross), a picture of your own (blood drops, a pair of fangs) at the height you choose, or a blank line. The picture is drawn in the manuscript, the preview, the paperback and the ebook. Your own characters print when the book's fonts have them, and asterisks print otherwise.
  • Under each chapter's title (Book ▸ Design ▸ Chapter openings ▸ Under the title): nothing, one of the typed ornaments, characters of your own, or the same as the scene breaks (their ornament, characters or picture, following any change to them).
  • Chapter art (Book ▸ Design ▸ Chapter art): a picture behind each chapter's first page with the chapter's words printed over it, and a picture on the left-hand page facing it. Choose them once for every chapter, and under them give any chapter its own or none. Only the chapters that differ are listed, so a book of hundreds of chapters takes no more room than one of three. Words on a picture sets the words over the art dark or light. A chapter with a facing picture opens on a right-hand page, with a blank page before it where needed. The art reaches the paper's edge when the paperback is exported With bleed. In the ebook each picture is a page of its own before its chapter. A picture for a single chapter needs the server updated to this version.
  • Book preview (the eye button, Manuscript): every page exactly as it will print, with facing pages on a desk, page 1 alone on the right, chapter jump, zoom, and "Edit here" from any line. On a wide window it docks beside the manuscript and re-sets live as you write (only the chapter you're editing is re-set). On a phone it's one page at a time with swipes.
  • Book exports (Export menu, Manuscript): the Paperback PDF is typeset with optimal line breaking, hyphenation, no widows or orphans, real small caps, drop caps, running heads and roman/arabic page numbers. Fonts are embedded, the trim box is set, and it is ready for KDP and IngramSpark. For a printer that asks for it, the paperback's dialog can make the file as PDF/X-1a (black ink only, so pictures are made gray) and with bleed (each page 0.125 in larger at the top, the bottom and the outside edge, with the page itself marked as the trim). The pages are set the same either way. The Booklet PDF is for printing and folding at home. The Ebook (EPUB) is for Kindle, Apple Books, Kobo and Google Play. With Embed the book's fonts on, the body and heading fonts' files go inside it (the app's own or the library's, each script its own file), so readers that honor a publisher's fonts show the book in them. The Submission manuscript (Standard Manuscript Format) exports as PDF or Word. Each shows a preflight list first (missing ISBN digits, glyphs a font lacks, a too-short book), each item with a "Show" button.
  • Book cover (Book inspector ▸ Cover, Manuscript): the cover is kept with the manuscript. The front cover is the ebook's cover and goes into every EPUB export. For the paperback, add a back and a spine to go with it, or the whole cover in one picture as a designer sends it. The cover sheet is sized from the trim, the paper and the page count the layout engine counts, and each picture is checked for sharpness and shape. If the page count changes after the artwork went in, it says by how much the spine has moved. The sizes follow Amazon KDP's rules.
  • Cover editor (Cover tab ▸ Open the cover editor, or Export ▸ Paperback cover): the sheet is drawn as it prints, whole or one part at a time (the spine lies on its side so it can be read). Click a part to give it a picture or a paper color, enlarge the picture and drag it to choose what shows. Add a title, a subtitle, an author, a spine line and the back's blurb, starting from the book's own details, and set each one's font, size, color, alignment, letter and line spacing, with an optional band of color behind it. Drag words into place, pull their sides to set the width, nudge with the arrow keys, undo and redo. The words are real text set by the book's own typesetting engine, so the screen, the paperback cover PDF and the ebook's cover agree. The editor also saves a template at the exact size with the printer's guides (what is trimmed off, the margin to keep clear, the spine, the barcode) for artwork made elsewhere.
  • Fonts and commercial use: every font that comes with the app and every family in the online library is under the SIL Open Font License, Apache 2.0 or the Ubuntu Font Licence. Each allows commercial use, so a book set in them can be sold and the fonts embedded in its PDF or ebook. A book's design and its cover offer only these. Documents can also use the fonts installed on your device. Those carry their makers' own licenses, which the app can't read, and the font list says so.
  • Version history: revisions are recorded at the seams of the work. A pause of half a minute closes one off, and a long unbroken stretch is cut on its own after five minutes. There are also named drafts you save yourself. Any revision can be previewed (in the document's own format, or as changes against the revision before it) and restored. Restoring saves a "Before restoring" version first and keeps comments on untouched paragraphs.
  • Find and replace (Ctrl+F, and Ctrl+H to replace, or ⌥⌘F on a Mac): finds phrases even across italics, curly and straight quotes match each other, and match case and whole words are a click away. Ctrl+G or F3 step through matches. Replace keeps the formatting and Replace All is a single undo. Opening it never moves your cursor until you step to a match. Footnotes and endnotes are searched too. A match inside a note marks the note's number, the count says "in a note", and Replace changes the note's words like any others.
  • Go to (click the page count, or Ctrl+Alt+G): type a page number, +2 or -1 for pages from here, or some words of a heading. The pages are the printed ones. A book's manuscript goes by headings.
  • Back to where you last wrote (Shift+F5): the caret returns to the last place you changed the text. Press again for the place before, up to four. What other people change makes no place of yours.
  • Select more (F8, or ⌥⌘↑ on a Mac): the word at the caret, then its sentence, then its paragraph, then everything. Shift+F8 (⌥⌘↓ on a Mac) goes back a step.
  • Formatting marks (Home ▸ More paragraph formatting ▸ Show ¶) now show a dot for each space, an arrow for a tab, a ring for a no-break space, a bent arrow where a line is broken by hand and a hook at an optional hyphen, beside the mark at each paragraph's end. They are drawn over the text and are never part of it.
  • Dark paper (the zoom menu, Settings ▸ Writing, or the moon on a writing session's page): the page dark with light text, for the screen only, in documents and writing sessions alike. Colored text is made lighter where it could not otherwise be read. Print, PDF and Word files are unchanged.
  • Formatting here (the top of the sliders panel's Text tab, or Ctrl+K "formatting here"): the style of the text at the caret, what its paragraph was given on top of that, and what the words themselves were given, each with a Clear of its own.
  • Move a section from the outline: in a Document the outline's Reorder moves a heading with everything under it, up or down past its neighbors, and takes it up or down a level with the headings beneath it. Drag a heading to do the same by hand.
  • Show only this chapter (the button at the end of a row in the outline, or Ctrl+K): the page shows one chapter, from its heading to the next. It is the same document, with the rest hidden and not taken out, so co-writers and exports see the whole text. A bar over the page names the chapter and steps to the one before or after. The caret keeps to the chapter, Select All selects it, Find looks only in it and the word count is its own. Going somewhere else from outside the text (Go to, the outline, a comment, the Check panel) shows the chapter that place is in. Show everything brings the whole text back. In a Document it shows a heading's section, in a screenplay an act.
  • Page count beside the word count: in a manuscript, the pages of the paperback as it is typeset with its front matter, the same count as the preview. Elsewhere it is the pages of the sheet you write on.
  • Spellcheck underlines as you type and uses the dictionary of the text's language, so a document set to German is checked as German on a device that speaks English.
  • Zoom the page you are writing on, from 50% to 300%. The figure beside the word count opens the sizes, with Page width and Whole page to fit the window. Ctrl with the wheel and a pinch on a trackpad change it too. The choice is this device's and holds for every document. Zooming moves no line of text: the page is laid out once and drawn larger or smaller. The print preview has a zoom of its own. Phones have none, since the page already fits the screen there.
  • Focus mode (the target button, Esc leaves): the app chrome steps away and it's just you and the page. Optional typewriter scrolling keeps the line you're writing vertically centered.
  • Reopen where you left off: each document remembers its cursor and scroll position on this device.
  • Word count: on every screen, including phones. With text selected it reads "120 of 12,345 words". Click it to set a word-count goal. Progress rides under the count as a thin bar, per document, on this device.
  • The ribbon (Document format): four Word-style tabs of controls. Home has undo, the format painter, styles (below), and font and size for the words you select (Larger/Smaller, Ctrl+Shift+. and Ctrl+Shift+, step Word's ladder, and the inspector takes any size). It has bold, italic, underline, strikethrough, subscript and superscript (Ctrl+= and Ctrl+Shift+=), text color and highlight, inline code, change case and clear formatting. It has bulleted lists with a choice of marker, numbered lists with a numbering style (1, a, A, i, I, or legal 1.1.1) and restart or continue, checklists (Ctrl+Shift+9) with boxes you tick, and indent/outdent (of a list item, or of a paragraph, half an inch at a time). It also has alignment (Ctrl+Shift with L, E, R or J), line and paragraph spacing, paragraph shading and borders (a box, a line above or below, or a bar at the left, where neighbors with the same border are one box), sort A to Z, and show paragraph marks (¶). Insert has page break (Ctrl+Enter), column break and section break, and a table (pick its size from a grid). It has link (Ctrl+Shift+K: an address, and the words to show when nothing is selected), picture, quote, code block and divider. It also has footnote, endnote and a table of contents, a symbol palette, and today's date. Layout has margins (Normal, Narrow, Moderate, Wide, or custom), paper and orientation, columns, sections, hyphenation, line numbers, headers and footers. A Table tab appears while the caret is in a table, and a Picture tab while a picture is selected. Review has Editing, Suggesting or Viewing, the suggested changes (next, previous, accept, reject), comments, Check, history, compare, find, read aloud, preview and print. Each tab's row holds only what is used every day. The rest is one click away under a More (⋯) button beside it: More text formatting (strikethrough, sub- and superscript, inline code, larger/smaller, change case, copy formatting, clear formatting), More paragraph formatting (shading, borders, sort, show ¶), and a More on Insert, Layout and Review. A More button is lit while something inside it is switched on. Alignment is one menu. A list's style menu shows while the caret is in a list. Accept and Reject show when there are suggestions to decide. Groups that don't fit the window fold into a ⋯ menu at the end of the row. On a phone the ribbon is one row that scrolls, and its menus open as sheets. While the phone's keyboard is up for the text, the row sits on the keyboard under your thumbs. It goes back to the top when you type somewhere else (the title, a comment, a search). The scripted formats keep their own toolbars (which ride on the keyboard the same way) and professional layouts.
  • Selection menu: select words with the mouse and a small menu appears above them. It has bold, italic, underline, text color, highlight, link, comment and clear formatting, so the commonest changes need no trip to the ribbon. It never takes the caret out of the text. Escape, or clicking anywhere, puts it away, and it follows the selection as the page scrolls (and goes when the selection scrolls out of view). Its last button (the sliders) chooses what it holds: tick strikethrough, sub- and superscript or inline code in, tick anything out, or switch the menu off. Settings ▸ Writing has the same choices (and switches it back on). (On a touch screen the system's own selection menu is left alone and the toolbar sits on the keyboard instead.)
  • Links: put the caret on a link and a small panel beside it says where it goes, with Open, Copy, Edit (its address and its words) and Remove. A click in the text only places the caret, so a link is never followed by accident. Ctrl-click follows it, and in a document you can only read, any click does. Links open in your browser (or mail program) and never inside the app. Escape puts the panel away.
  • Styles (Document format): Normal, No Spacing, Title, Subtitle, Heading 1–6, Quote and Caption, each shown in its own look, plus styles of your own. Modify… changes what a style looks like everywhere it's used (font, size, color, alignment, spacing, indents). Update to match selection makes the style look like the paragraph you formatted by hand. New style from selection names that look. Styles are saved in the document, so collaborators, the PDF and the Word file all have them. Normal is the document's body text (Document tab of the inspector). After a title, subtitle or caption, Enter goes back to body text.
  • Format painter: click it, then select other text. It takes the formatting of the words you were on (and whole paragraphs take the paragraph's alignment, spacing and indents too). Escape puts it down.
  • Pasting from Word, Google Docs and the web (Document format) matches this document by default: the words come in with their structure (paragraphs, headings, lists, tables, links) and their emphasis (bold, italics, underline, sub/superscripts), and leave their look behind: typeface, size, color, highlight, alignment, indents and spacing are this document's. A small strip under the paste says so and offers Keep its styling. One click brings the colors, highlights, alignment and differing fonts and sizes back (and Match this document switches again). The choice stays until you type the next thing, and Escape or the × settles it sooner. Word's typed-out bullets become real lists. What you copy inside the app pastes exactly as it was. Ctrl+Shift+V still pastes plain text.
  • Tables (Document format): Insert ▸ Table, then type. Tab moves to the next cell and adds a row after the last. The Table tab adds and deletes rows and columns, merges and splits cells, marks a header row (set again at the top of each page the table runs onto), chooses which lines are drawn (all, between rows, a box, none) and shades cells. Drag a column's right edge to resize it, or use Narrower/Wider (on a touch screen). Even widths shares the page out equally. A table from Word, a spreadsheet or a web page pastes as a table. A table is never wider than the sheet: where the window is narrower than the paper, its columns keep their proportions. A cell can have borders of its own (a colored rule down the side of a column, say). They come in from Word and go back out, though there is no control for drawing them here yet. In a table with no lines, faint guides show where the cells are while the caret is in it or the pointer over it, and nowhere else. On paper a table breaks between rows. A row that won't fit what is left of a page is cut across the pages, each cell going on at the top of the next, and page view shows the same break.
  • Footnotes and endnotes (Document format): Insert ▸ Footnote puts a numbered reference where the caret is and opens the note under it to write in (bold, italic and underline work, and Enter or Esc goes back to the text). Notes are numbered by themselves (1, 2, 3 for footnotes and i, ii, iii for endnotes) and renumber as you move text about. A note travels with its sentence when you cut and paste, and several people can write in one at once. In preview, print and PDF a footnote sits at the foot of the page that cites it (running over to the next page if it must) and endnotes close the document. Clicking a reference in the PDF goes to its note. A note's words aren't counted in the word count. While you write, every note is also listed under the document's last page (Footnotes and Endnotes, each with its number). Click a note's words to change them right there, or its number to go to the sentence that cites it.
  • Table of contents (Document format): Insert ▸ Contents lists the document's headings and keeps itself up to date. Click an entry to go to it. On paper each entry has its page number and a row of dots, and in the PDF each is a link.
  • Sections (Document format): Layout ▸ Section starts a new section at the paragraph the caret is in. A section begins on a new page and can have its own paper, orientation, margins, number of columns, and page numbering that starts again. What it doesn't set is as the document. A landscape page in the middle of a portrait document is a section. A column break (Insert ▸ breaks) sends what follows to the top of the next column.
  • Pictures (Document format): drag a corner handle to resize and the proportions are kept. The Picture tab sets one at the left, in the middle or at the right, gives it a border (a color and a thickness) and rounded corners (up to a circle), an exact size in inches, a caption that stays under it on the page, and a description for someone who can't see it. A picture is a block of its own between the other blocks, with no anchors or wrapping to go wrong. To move it, drag the handle on its corner to where it should go (a line shows the place), or use Up and Down on the Picture tab or Alt+↑ and Alt+↓. Borders and corners come in from Word and go back out.
  • Document inspector (the sliders button, Document format): settings saved in the document, so collaborators see the same layout. Text tab: the selection's font and size (any size), alignment, line spacing (a multiple, or exactly / at least so many points), space before and after, indents (left, right, first line or hanging), keep with next, keep lines together, start on a new page, plus a drop cap for chapter openers. Document tab (the page setup): paper (US Letter, Legal, Tabloid, Executive, Half Letter, US Trade 6×9, A3, A4, A5, B5, or a custom size. A new Document starts on the paper of where you are: US Letter in the countries that print on it and A4 everywhere else. The choice is kept with the document, so everyone who opens it sees the same sheet) and orientation, margins, a binding gutter and facing pages (mirrored margins), and one to three columns. Headers and footers have a left, center and right slot each and live fields (page number, page count, title, author, date), and are optionally different on the first page and on even pages. Page numbers are in 1/i/I/a/A style from any starting number. Body-text defaults cover font, size, line spacing, paragraph spacing and first-line indent (indent + 0 pt spacing gives the classic book look). Hyphenation and line numbers are here too.
  • Preview (the eye button, every format): every page exactly as it will print, in the same layout the PDF and the printer get. A screenplay, a stage play and a poem preview on their format's own page. Docked beside the text on a wide window, it follows your typing and shows Page 3 of 12. Click any line for Edit here.
  • Print (Ctrl+P, or Export ▸ Print…): the document is typeset to a PDF and handed to the system's print panel. On a Mac that is the native one, on an iPhone the system print sheet, and in a browser or on Windows the browser's own. Where that isn't possible it opens as a PDF you can print from. "As a booklet" imposes the pages for folding.
  • Screenplay, stage play and poetry pages: their PDFs, printouts and previews are typeset by the same engine as a Document's, on the format's own page. A screenplay is set in twelve-point Courier, six lines to the inch, with scene headings, action, character names, parentheticals, dialogue and transitions at their standard measures and the page number at the top right from page two. A stage play is set likewise and opens with its title. A poem is set in a book face with long lines turned over and set in, and its title at the top. Any alphabet prints (the old typewriter pages turned everything but Latin into question marks), a character's name is never left at the foot of a page without the speech, and bold, italic and underline are kept.
  • PDF (Document format): typeset in real typefaces with proper line breaking. It has mixed fonts and sizes, bold and italic, underline and strikethrough, lists, quotes, code, pictures, links you can click, page breaks, an outline of your headings, headers, footers and page numbers, with no line left alone at the top or foot of a page and headings kept with their text. Every font prints as itself where it can (see Fonts). Where it can't, the export says what stood in.
  • Export PDF (Export ▸ Export PDF…, every format but Manuscript): a small dialog with the choices for the file, remembered on the device. Pictures go in for a Screen (150 pixels an inch, a file small enough to send), for Print (300) or as the Original files. A picture is sized by how large it prints, so one shown small costs little. A JPEG that is already good enough goes in untouched and is never encoded a second time. A photograph taken on its side stays the right way up, a see-through picture stays see-through, and a screenshot stays sharp. What a camera wrote beside a photograph (where and when it was taken) is not copied into the PDF. Tags for screen readers writes the document's structure into the file: headings, paragraphs, lists, tables with their header cells, notes, links and each picture's description, in reading order. Running heads and page numbers are marked as decoration. The export says how many pictures have no description yet. Archive copy (PDF/A) makes a PDF/A-2b file, meant to open the same way in many years: every font is inside it, and it says what its colors mean. Keep an editable copy inside puts the document itself into the PDF beside its pages (see Import). Details are the subject, keywords and language. They are kept with the document, travel in a Word export too, and are what search and screen readers look at.
  • What every PDF says of itself: its title, author, subject, keywords and language, when it was made, and its page labels, so a reader's "page iv" or "page 12" is the page that prints that number.
  • Fonts (Document format): the font button opens one list of everything you can set a paragraph (or the whole body) in, each drawn in itself and searched together. The list has the fonts in this document, your recent ones, those with the app, the fonts on this computer (desktop), and an online library of nearly 2,000 open-licensed families (the Google Fonts collection), filtered by kind: serif, sans, mono, display, script. A library font downloads the first time a document uses it and is kept on the device. Other scripts of a family (Cyrillic, Greek, Vietnamese…) download only if your text has them. The same font files draw the editor, the preview, the PDF and the printout, so they agree. Installed fonts are embedded in a PDF only when the font's own license setting allows it, and are never uploaded. When a document names a font this device doesn't have (a collaborator's, or one from a Word file) it's shown and printed in the closest stand-in: a same-metrics family where one exists (Gelasio for Georgia, Tinos for Times New Roman, Arimo for Arial, Carlito for Calibri), so lines and pages fall the same. A line under the ribbon says so and offers Replace…. Settings ▸ Fonts has the switch for the online library, the downloaded fonts (and a way to remove them), and the licenses.
  • Word (.docx) in and out, including fonts. A Word file's fonts are matched to the app's own, to those installed here, then to the online library, and otherwise kept by name. Exports carry real family names and tell Word what to fall back to. Both directions keep headings 1–6, real nested and numbered lists (bullet styles, legal numbering and checklists included), bold/italic/underline/strikethrough, text color, highlight, sub/superscript and a font or size on part of a paragraph, links, pictures with their sizes, alignment, spacing, indents, line-height rules, borders and shading, keep-with-next, quotes, code, dividers, page breaks and drop caps. They also keep tables (widths, merged cells, header row, shading), footnotes and endnotes, a table of contents (a real field Word can update), picture placement and captions, sections and column breaks, plus styles (a Word file's Title, headings and the writer's own styles come in as the document's styles, looking as they did, and yours go out as real Word styles) and the page setup, headers, footers and page numbers. What Word shows is what the document says, and a file written by subScribe reads back the same.
  • Page view (Document and the scripted formats): the page toggle shows the sheet as its printed pages, with gaps between them, page numbers on the left and the count in the word-count bar. The pages are the printed ones: the text breaks where the PDF's pages do, in the middle of a paragraph or between the rows of a table, and the count is the PDF's. It keeps up as you type. The gaps between pages hold still and the text moves across them. Lines within a page are still the screen's own, so a page whose text needs much more room than the paper is made taller. Preview shows the page line for line. Settings ▸ Writing has a switch back to the simpler view that only moves whole paragraphs.
  • Outline: parts, chapters and sections with their labels and word counts. Drag to reorder chapters (or use Reorder on a phone). One undo puts them back.
  • Side panels at the width you want: drag the edge between the text and any panel beside it (the outline, comments, history, the inspector, the docked preview, the plan, Check) to make the panel wider or narrower. The arrow keys move the edge too, and a double click puts the panel back as it was. Each kind of panel keeps its own width on this device. The text always keeps room of its own.
  • Story plan: plan a book (or a script, or a Document) beside its text, from the Plan button in the header, the outline, or Ctrl+Alt+P. It is kept in the document, so everyone who writes it sees the same plan as it changes, and it goes with the document when it is published or copied.
  • Board: the chapters across the top (a script's scenes, a Document's sections), with their words and parts above them, and plot lines down the side. Cards in the cells hold a title, notes, a color and who is in them. Drag a card to another chapter or plot line (on a phone or tablet, hold it still first), or use its menu's Move to…. Notes not yet in a chapter wait under the board. A column's menu shows the chapter in the text or on its own, renames it, or adds a chapter after it, and those change the text itself. A row can show who appears in each chapter. List puts the chapters down the page, and phones always show it that way.
  • Cards follow their chapter: through a new title, a move in the outline, publishing and copies. A chapter deleted leaves its cards under the board with its name, and Undo brings them back.
  • Cast and world: people, places and things, each with other names it goes by, notes and a color. Each shows the chapters it appears in, found in the text by its names, and its mentions can be stepped through one by one. A name written with a capital is only found with one, and a name that is also a common word (Will, May) is passed over at the start of a sentence.
  • Timeline: events in the order they happen in the story, each with when it happens and the chapter that tells it. Beside them is the order the book tells them in, with lines between the two, so a flashback shows where it crosses. Events with dates can be sorted by them.
  • Beside the text: the side panel shows the cards of the chapter you are writing in and who is in it.
  • Copy the plan as Markdown from the view menu. Export all and the export of everything on this device write each plan beside its document.
  • A chapter's tie to its cards is new in this version. A book with cards in its chapters can't be opened by older versions until they update, and its server must be updated first. Until then the plan works without chapters.
  • Organizer: cards show each document's opening words (or, when searching, the passage that matched) and "Manuscript · 98,412 words · Oct 3". A book's card also shows its cover, its description in place of the opening words and its tags. A Document's card shows its subject and keywords the same way. The card is the same for everyone the document is shared with, in the share card in chat and in Home's recent list too. The cover on the card is a small picture the app draws of the front as designed, with its words on it. Sort by modified, created or name (in natural order: Chapter 2 before Chapter 10). The sort is remembered on this device. Every ⋯ menu works from the keyboard (Shift+F10 too): Rename, Duplicate, Move, Share, Version history, Delete.
  • Recently Deleted: deleting is undoable. The toast has Undo, and deleted documents (from the server and from this device) wait in Recently Deleted for 30 days, with their history, before they're gone for good.
  • Combine into one manuscript: select documents (Ctrl-click, Shift-click, or Select on a phone) or use a folder's menu. Put them in order, name the book, and each becomes a chapter. The originals stay.
  • Folders: organize your documents. Move documents between folders, and rename or delete a folder (its documents stay put). Documents on this device have folders of their own, which are kept on the device. New folder asks where when there's a choice, and a folder's card says on this device. New documents, imports and combined manuscripts made with such a folder open land in it, and a document restored from Recently Deleted goes back to the folder it was in.
  • Search: full-text. It matches titles and the words of every document (never the editor's internal markup) and shows the matching passage. Documents on this device are searched the same way, with or without a server: the first search after an update reads the documents once, and after that it's instant.
  • Suggested changes (track changes): switch to Suggesting and what you type, delete or paste is proposed instead of done. Additions are underlined in your color, deletions are struck through, a new or removed paragraph break is marked with a ¶, and a table, picture, divider or footnote is outlined. Each says who and when. Anyone who can edit accepts or rejects them one at a time or all at once, from the Review tab or from the Changes list beside the comments (where a change can also be commented on). Taking back your own suggestion just removes it. The mode is yours alone, so one person can suggest while another edits. Viewing shows the text as it would read with everything accepted and keeps it locked. Formatting isn't tracked: bold, styles, lists and a table's rows and columns change directly. Printing, PDFs, books and the word count use the text with suggestions accepted. Suggested changes travel to and from Word as its tracked changes. Export Word (clean copy) leaves them out. Works in every format (the mode switch is in the Changes list, and a Document also has it on its ribbon). Like the other newer formatting it needs an up-to-date server. Where the server is older, Suggesting isn't offered.
  • Read aloud: the text is spoken by your system's voice a sentence at a time, with the sentence highlighted as it's read. It reads from the caret, with pause, back and on a sentence, speed and a choice of voice. Nothing is sent anywhere. Struck (suggested-deleted) words and footnotes aren't read.
  • Check: a reading of the whole text for slips and for things written two ways. Open it from the Review tab, the header button or F7, in any kind of document. Each finding lists the places it occurs with the words round them. Choosing a place selects it in the text. Fix changes one place and Fix all changes every one as a single step you can undo. In Suggesting mode each fix is a suggestion. Where the text does a thing two ways, you say which is meant and the others are listed to change. It looks for:
  • words spelled the American way in some places and the British way in others (color, colour). Choose American, British, British with -ize, Canadian or Australian in Rules and every spelling outside it is listed. Names are left alone, and so are spellings that are also another word (check, tire, story)
  • one word spelled two ways (judgement, judgment) and -ise mixed with -ize
  • a word with its accent in some places and without in others (café, cafe)
  • slips a spelling checker passes because each word is a word (could of, sneak peak, its a). How people talk is left alone inside quote marks
  • a phrase hyphenated in some places and open in others (a short-term plan, a short term plan). "In the short term" is a different case and is not weighed against it. Also one word hyphenated and closed (co-operate, cooperate)
  • a range of numbers with a hyphen and with an en dash, and the dash in a sentence drawn more than one way
  • numbers in a sentence as words in some places and figures in others, and twenty-one with and without its hyphen. Labels, measures, times, dates and money are not counted
  • a word with a capital in some places and without in others (the President, the president). The first word of a sentence, a word beside another capital, and words that are also names (Will, Rose) are not weighed
  • headings of one kind capitalised in more than one way
  • two names so alike that one may be the other misspelled (Katherine, Kathryn). Two names that keep turning up together are two people and are left alone
  • the comma before "and" in a list, with it in some places and without in others. Rules can ask for always or never
  • the items of one list ending or beginning differently
  • a title or a name in italic in some places and plain in others
  • an abbreviation written two ways (NASA, Nasa). In a Document it also looks at definitions: explained twice, given two meanings, used before it is explained, never explained, explained and never used
  • the same word twice in a row (the the)
  • brackets and quote marks opened and not closed. A speech that runs over several paragraphs and closes only the last is left alone
  • straight quote marks mixed with curly ones, and the same for apostrophes
  • speech in double marks in some places and single in others
  • two spaces in a row, a space before punctuation or inside brackets, no space after a comma, spaces at the end of a paragraph
  • what a draft leaves behind: highlighting, TK, TODO, XXX, notes in square brackets, open comments and suggested changes

The panel's Rules tab holds the house style, in two places. This document has rules of its own that travel with it and that every co-writer shares. All my documents has your own rules for everything you open on this device. The document has the last word, and it can ask that only its own rules be used so that everyone sees the same findings. A rule can: - switch a check on or off - write one word or phrase for another (per cent as percent), fixed in the capitals it was written in - point a word out each time it is used, with a note of your own - say that a word or phrase is never to be flagged

A place you skip is remembered in the document by the words at that place, so it stays skipped when the text moves and comes back if those words change. A whole finding can be set aside for now, for this document, or in all your documents. Rules save to a file and load from one, and your own are part of "Export everything on this device".

The Reports tab (in a manuscript and in a Document) shows what the text is like. Nothing in it is a fault to fix. Each report says what was counted, and choosing a row shows it in the text. - Overview: words, sentences, paragraphs, the average sentence, a reading grade and the share of the words inside quote marks - Chapter by chapter: one row a chapter with its length, its share of speech and its average sentence. A chapter is marked where a number stands well apart from the other chapters. Sort the table, or choose a row to go there and make every report about that chapter - Sentences: a chart of every sentence's length in order (click a bar to go to it), how sentences begin, sentences in a row that open with the same word, runs all of one length, long sentences and long paragraphs. Speech is left out of the patterns - Repeats: the same word again within about thirty words, phrases said again and again, the words used most, and a word one chapter leans on far more than the rest of the book. Names and small words are never counted. Say This is fine to a word and it is left alone, in this document or in all of yours - Word choice (English): likely passive verbs, words in -ly listed by word, filter words (she saw, he felt), feelings told (felt angry), padding words (very, really, just), worn phrases, words that say it twice (nodded his head, stood up), wordy phrases (in order to), and every use of words that are easily taken for one another (affect, effect). Where there is a plain fix it is offered, one place or all of them, and made as a suggestion in Suggesting mode. In fiction, speech is left out unless you ask for it - Dialogue (English): how much of the text is speech, the tags that carry it (said and asked, other verbs, actions such as smiled), tags with a word in -ly, long exchanges where nobody is named, and punctuation amiss between speech and tag, each with its fix. A sentence of its own after speech is a beat and is left alone - Pace: the text in stretches of about 300 words, each drawn as quick, steady or slow-leaning, and a list of long stretches with little or no speech. It is told from the shape of the text alone (how much is speech, how long the sentences and paragraphs are, how often it looks back or goes inside a head), so a stretch is named for what was measured and never called a fault - Senses (English): words that name sight, sound, touch, smell and taste, counted outside speech. There is no right balance, so none is given

A report covers the whole text, the chapter the caret is in, or the paragraphs you have selected. An outside figure is shown only where a published source gives one (the Plain English Campaign's 15 to 20 words for the average sentence of writing that explains). For fiction the reports compare your chapters with each other.

The checks that depend on English (spelling, accents, slips, hyphens, capitals, headings, numbers, the serial comma) run only for English text, and the panel says how many did not run. Of the reports, Word choice, Dialogue and Senses read English. Overview, Chapter by chapter, Sentences, Repeats and Pace run in any language with spaces between its words, and Repeats knows the small words of French, German, Spanish, Italian, Dutch and Portuguese. The table of which English uses which spelling comes from VarCon, part of the SCOWL project, used under its license. Its notice is shown at the foot of the Rules tab.

The arrows beside Close make the panel about twice as wide and back, and its edge can be dragged to any width between. Wide enough, the charts have more room, the reports sit two abreast and the chapter table shows more of each chapter. The width is kept on this device.

Nothing is sent anywhere, and nothing runs while the panel is closed. It reads the text again after each pause in your writing and never moves your selection. On a phone it opens as a sheet under the page and folds to a bar with previous, next and Fix while you step through the places. - Find by pattern: the pattern button in Find and replace makes the words to find a regular expression. In the replacement, $1 is what the first pair of brackets found. Replace all is still one undo. - Compare: any saved version beside the text as it is now, or the document beside a Word file. Paragraphs are aligned, and what was taken out is marked on one side and what was put in on the other, word by word. On a phone both are in one column. Jump between differences, show only what changed or the whole text, swap the sides. - Feedback threads: select text and leave a comment anchored to it. Threads track the text even as it moves, resolve/reopen, reply inline, and list in the order they appear in the text. Click a highlighted passage to open its thread. The highlighter button in the panel header shows or hides the anchor highlights. On-device documents have the same panel as Notes: comments to your future editing self. - Edit a PDF (pick a PDF with Import and choose Edit this PDF): the PDF opens as the PDF it is, to read, search and change. Pages: move them by dragging in the list, turn them, copy them, take them out, add an empty page or the pages of another PDF, and save some pages as a PDF of their own. Mark up: highlight, underline or strike through selected text, add notes, draw, put a box or an arrow, and type words on the page. Forms are filled in on the page. Sign by drawing your signature or using a picture of it. That is a signature as ink on paper is. It is not a certificate. Change text (Beta) lets you click a piece of the page's own text and retype it. Every step can be undone. Your file is never written over: Save a copy writes the result where you say, and the apps can also keep a PDF on the device to come back to, listed in Documents and in Home. A PDF with a password asks for it. One whose owner does not allow changes says so and keeps those tools off. - Import: PDF, DOCX, RTF, Pages, TXT, and Markdown files convert into new documents. Several can be imported at once, and they go into the folder you have open. PDF import reads the pages back into a document: paragraphs (a word broken at a line's end is put together again), headings, lists, tables, links, footnotes, and pictures with what is said under them, in the order the page is read even where it is set in columns. The paper, the margins, the columns, the font and size of the text, its colors, and what was printed at the head and foot of every page become the document's own page setup, header and footer. A tagged PDF (from Word, Google Docs, InDesign or this app) is read by its tags and not guessed at. A PDF with a password asks for it. A scanned PDF asks whether to read the text out of its pages or keep them as pictures. Reading uses the system's own text recognizer on a Mac, an iPhone or an iPad. It runs on the device and sends nothing anywhere. Other devices keep a scan's pages as pictures. A long import shows how far it is and can be canceled. A note afterwards says what was left out, such as drawings and text set at an angle. A PDF exported from here with Keep an editable copy inside holds the whole document, with its styles, page setup and pictures. Importing it asks whether to open the document or read the pages. Suggested changes were saved as accepted, and comments and history are not in it. Anyone who has such a PDF can open the document the same way. Word import keeps headings, lists, links, pictures, fonts, spacing and the page setup (see above), and says what it couldn't keep yet: text boxes arrive as their text, and a section break that doesn't start a page is left out. Word's tracked changes come in as suggested changes, under the names Word has for them. A Word manuscript arrives as a Manuscript with its chapters detected. - Booklet export (Document and scripted formats): beside the straight PDF, "Export booklet PDF" imposes pages for saddle stitching, with two per landscape sheet in fold order. Print two-sided flipping on the short edge, fold the stack in half, and it reads 1…N. - Saving files: on a Mac or PC every export opens a Save dialog (the toast after offers Show in Finder). On iPhone and iPad exports land in Files ▸ On My iPhone ▸ subScribe, with a Share… button. In the browser they download. - Export all: back up every document you can see (server documents and the ones on this device) as Markdown + JSON. Your writing is never locked in. - Sharing: private by default. The owner grants read or write access to individual friends or to a whole group (which also posts a card in that group's chat) or shares an entire folder at once. - Canvas references: send a whole document or just a selection to a canvas board as a live-updating excerpt card. - Documents on this device: the "On this device" section holds single-user documents stored on your computer and never on the server. They have the same editor, formats, outline, find, word count, version history, book preview and exports. Their revisions are kept on this device too, beside the document, so history works with no server and no connection. Publish to a server… (a document's ⋯ menu, or the share button in its header) publishes a copy, optionally shared with a friend or group in one step. The local original stays yours and does not live-sync to the published copy. If you aren't connected it opens the connect sheet first and carries on once you are. Without a server, Import creates documents here too. One document, not a copy for each group. Publishing remembers the server copy it made. Publish again, to share with another group, and the dialog opens on that same document, says who has it and shares it with more people. A document of the same name already on the server is offered before another copy is made. Keep only the server copy moves the one on this device to Recently Deleted, so the server copy is the only one. A server document is on every device you sign in on. Your own documents' cards say who they are shared with. The reverse works as well: any server document's menu → Save a copy to this device. Pictures work here as they do on a server: they're kept beside the document on this device, travel with it to Recently Deleted and back, are embedded in its PDF and Word exports, and are uploaded with it when you publish.

Chat

Groups work like small Discord servers: channels, invites, roles.

  • Invites two ways: admins generate expiring invite codes, or set a permanent join code (gear next to the group name). A join code is a memorable code like writers-club that works in the normal join box until an admin clears it.

  • Markdown: **bold**, *italic*, ~~strike~~, `code`, fenced ``` blocks, links.

  • @mentions and #channel references: @username pings and highlights, and #channel links and jumps.
  • Read state that follows you: what you've read syncs through the server, so unread counts agree across your devices. Opening a channel lands on a "New messages" divider where you left off, and messages are grouped under day separators (Today, Yesterday, …).
  • Typing indicators: a quiet "…is typing" line above the composer, in channels and DMs.
  • Search: the magnifier in the channel header searches the full text of every channel in the group. Click a hit and you land on that message in context, where it is highlighted.
  • Pinned messages: pin from a message's hover actions (your own, or any as an admin). The pin in the header counts them, lists them, and jumps.
  • Reactions can use the group's custom emotes as well as emoji.
  • Mute a channel: the bell on a channel row (or a right-click) silences its notifications and badges on this device. Unread shows as a quiet dot.
  • Custom emotes per group (admins add them via the emote picker). Type :name: anywhere.
  • Uploads: attach, paste, or drag & drop images.
  • Presence: online / busy / invisible, shown next to every name.
  • Profile cards: click anyone's name or avatar for their bio, accent color, and status. The card has a Message button for friends (straight into your conversation) and Add friend / Accept friend request for everyone else.
  • Moderation: authors and group admins can delete messages. Admins can kick members, rename and delete channels, and delete the whole group (gear next to the group name → Delete group). Deleting a group irreversibly erases its channels, messages, sessions, boards, and emotes for every member.
  • Slash commands: type / to see commands contributed by plugins.

Voice & video

Voice channels sit under a group's text channels. Media is a direct peer-to-peer mesh between members (DTLS-SRTP), and the server only relays connection setup. A call has mute, speaking indicators, and a persistent voice bar in the status bar, and it follows you around the app.

Turn on your camera or share your screen from the voice bar. Video appears in a floating panel you can drag, resize, and keep open while you work in Canvas or Documents. Click a tile to enlarge it. Peers' tiles show mute and screen-share badges. Screen sharing is available on Windows and Linux. macOS can watch shares but not send them (its webview has no screen capture), and camera video works everywhere.

Settings → Voice lets you pick input devices, adjust input gain and output volume, test your mic and camera, and choose video quality (360p default: the mesh sends a copy to every peer, so higher quality multiplies upload bandwidth).

Canvas

Shared storyboards per group: sticky notes, text, shapes (rectangle, ellipse, diamond, triangle, star or heart, with centered labels and colorable outlines, hollow or filled), colored frames, connectors, pasted or dropped images (drag a file onto the board and it lands under the cursor, and images are croppable from any edge, with rotate/flip), link cards with server-fetched previews, and live document references (a whole document or a selection, kept in sync as it's edited). Multi-select with shift-click or shift-drag. Moving a frame moves what's inside it. Note, frame and shape fills are translucent, so overlapping them layers the color instead of hiding what's underneath, and a selected note has an opacity slider of its own. There is a snap-to-grid toggle, and holding shift while dragging a resize handle keeps the element's proportions. Each element has its own text formatting (bold/italic/underline, size, alignment, bullets, four typefaces, text colors). Connectors run straight, at right angles or as a smooth curve, with per-side anchors, line weight and color, dashed styles, arrow/dot ends, and a label that rides the middle of the line. Selecting several elements offers a text color for all of them at once. Everything syncs live.

Lining things up: drag an element near a neighbor and smart guides appear. They are accent lines that show when edges or centers align, with a gentle snap (they outrank the dot grid on the axis they match). A multi-selection's toolbar has an align & distribute menu (left/center/right, top/middle/bottom, even out the gaps). Ctrl+G groups a selection so it selects and moves as one. Click again to pick a single element out of the group, Ctrl+Shift+G ungroups, and copies of a group stay grouped on their own.

Pages: a board is a stack of pages, listed above the canvas. Add one with the plus, double-click a tab to rename it, right-click for rename and delete (deleting takes its contents with it, and undo brings both back). Each page is its own space: what you draw, select, or fit to screen belongs to the page you're on. Peers' live cursors appear only on the page you share with them. A dot on a page tab (their accent color, or a count) shows someone is working over there.

Linking as you go: select an element and four arrows appear around it. Click one and you get a new element on that side, already connected and ready to type into. It is the quickest way to grow a diagram outward.

Sketches: the pencil in the toolbar opens a pad you can draw on with a mouse, trackpad, finger or stylus. Pick an ink color and nib, switch to the eraser to rub out strokes you've drawn, undo a stroke at a time, then insert it onto the board. Sketches stay editable: double-click one (or use the pencil on its toolbar) to reopen the pad and keep drawing. They're stored as strokes rather than as a picture, so they stay sharp at any zoom, resize cleanly, and work on a board with no server behind it. One person draws at a time: while someone has a sketch open, others trying to edit it are told who's in there instead of silently overwriting each other.

Copy and paste: Ctrl+C, X, V and D copy, cut, paste and duplicate the selection, and Ctrl+A selects everything on the board. A connector with both ends in the selection comes along and is re-pointed at the copies, so a pasted diagram stays wired the way you drew it. Copies go through the system clipboard, so they can be pasted into another board.

Right-click anything on the board for bring to front and send to back, cut, copy, paste and duplicate, lock (a locked element still selects, so you can unlock it, but won't move or resize), round corners, and delete. Everything applies to the whole selection. Right-clicking the board itself offers paste, select all, and fit to screen. Selecting one element also puts a corner-radius handle inside its top-left corner, opposite the scale handle: drag it away from the corner to round it off, back toward the corner to square it up.

Getting around: press F to frame the current selection (the view centers on it and zooms to fit) or, with nothing selected, to fit the whole board on screen.

Find and replace: Ctrl+F (or the search button in the board header) opens a bar that searches every note, text block, shape label and frame name. Enter and Shift-Enter step through the matches in reading order, selecting and centering each one, and you can replace the current match or all of them at once. A Replace All is a single undo step. Link cards, images, and document cards aren't searched: their text holds a URL, an attachment, or a document reference rather than prose.

Background: the palette button in the board header dresses the page you're on with a background color, or an image you upload (filled, fitted, stretched, tiled, or centered), and can turn the dot grid off. Each page keeps its own, everyone in the group sees it, and undo covers it like any other change.

Export & import boards: the download button in a board's header saves it as a .wfboard file. The file is self-contained and holds the board's pages, elements, connectors and pictures. Import one from the board list into a group or onto this device. The round trip is lossless, which also makes it the way to move boards between servers. The board list also imports FigJam and Freeform PDF exports (desktop app only). Since a PDF stores drawings rather than board structure, the import reconstructs it: pictures become image elements, small text cards become notes, larger cards become tinted frames holding their photos and text where they were laid out, and each imported page keeps the export's white paper so its black text reads as it did. The result is editable but approximate. Card colors map to the board palette, and freehand ink and decorations don't survive.

Boards on this device: the canvas list has an "On this device" section for boards stored on your computer and never on the server. These are the same boards with the same tools, kept in a file beside your local documents. They need no group and no connection, so they work in offline mode. That includes pictures: paste one and it's stored beside the board on your computer, and so are background images. The trade is that nobody else can see the board, there are no live cursors, and document cards (which point at server documents) stay on group boards. Publish to a server… (the share button in the board's header, or the board's ⋯ menu in the list) puts a copy (pictures included) in a group you choose. The board on this device stays, and later changes to it don't follow. Not connected, it opens the connect sheet first. The search field above the list finds boards by name and, for boards on this device, by the words on their notes, text, shapes and frames.

Notes

A local, Obsidian-compatible markdown vault with [[wiki-links]] and backlinks. Share a note snapshot to a friend over DM, and they can save it into their own vault.

  • Folders: the sidebar is a tree of folders, and they are real folders in the vault, as Obsidian has them, so a vault you open in Obsidian has the same folders. Make one with the folder button (inside the folder you last picked), and use a folder's ⋯ menu for a new note or folder in it, to rename it or to delete it. A deleted folder and everything in it wait in the vault's trash for 30 days, and the toast puts them back. A new note goes in the folder you last picked.
  • Moving a note: the folder beside a note's title is where it is. Choose another to move it there.
  • Links across folders: [[Name]] finds the note of that name in the same folder first, else the nearest one, as Obsidian does. [[Folder/Name]] names one exactly. Renaming or moving a note or a folder keeps every link to it working. A link to a note that isn't there makes it beside the note you clicked it in.

Getting around

  • Ctrl+K is the quick switcher: type a few letters and jump to any channel, group, document, board, note, or conversation, or run a command. While a document is open it also runs the editor's commands by name: type "table", "heading 2" or "page break" and press Enter. Each command shows its key beside it, the selection stays where it was, and the text has the focus again afterwards. Only what can be done just then is offered, so someone who can only read a document is offered Find and Print and nothing that changes the text.
  • Ctrl+/ shows a keyboard-shortcut cheat sheet, with every key of the document editor listed by what it is for.
  • The desktop window remembers its size and position between launches, and unread counts appear on the app's dock icon (macOS/Linux). The installed web app badges its icon too.
  • The desktop app quietly checks for updates a few seconds after launch and mentions a waiting one. Settings → Application shows the release notes before you install.

Notifications

Native OS notifications for direct messages, @mentions, new writing sessions, shared documents and notes, and friend requests. They are delivered only when you're not already looking at the relevant conversation.

Customization

  • Profile: avatar, banner image, accent color, bio, and status (Settings → Profile). Your accent tints your name in chat and fills the profile-card banner when no banner image is set.
  • Portable profile: save your look (display name, colors, bio, avatar, banner) on your computer and apply it to any server you join. Profiles are otherwise per-server. It's offered once when you register on a new server and available any time from Settings → Profile. It is never applied without asking. Usernames are per-server and aren't included. It is also who you are when no server is connected. Settings → Profile then edits it directly: name, accent, about, and a picture and banner chosen from this device (a large photo is scaled down first).
  • Groups: admins set a group icon and accent color (gear next to the group name). The active group's accent shows as a stripe on the app rail, and Canvas and Sessions show a chip naming the group you're working in. Share and send-to-canvas pickers preselect it too.

In the browser

Servers can serve the app at /. Open the server's address in any browser (phones included) for chat, documents, canvas, sessions, and voice in a responsive layout. Use your browser's Add to Home Screen to install it. Once installed it opens full-screen with its own icon, like a native app. Desktop-only features stay on the desktop: the notes vault, on-device documents and boards, the portable profile, hosting, and plugins.

On iPhone and iPad

subScribe also builds as a native iOS app (the same codebase through Tauri). Beyond the responsive layout the web client already has, the native app adds what a browser can't:

  • Real notifications through iOS, with permission asked the first time something wants your attention.
  • Haptics: a tap when a message sends, a success thump when you accept a friend request, a warning buzz when a sketch is locked by someone else drawing on it.
  • Voice keeps going when you lock the phone or switch apps, like a calls app.
  • Liquid-glass chrome: menus, toolbars, dialogs, and the phone's slide-over panels are translucent materials that blur the content moving beneath them.
  • The keyboard pushes the composer up instead of covering it, the app never rubber-bands off its edges, and launch fades straight into the app's own dark background.
  • An open document has the whole screen on a phone. The tab bar and the status strip step aside while you write, and the editor's header keeps to one row: the way back, the title, whether your words are saved and More. More holds the format and every tool with its name beside it. The page's margins are a phone's, so the words get the width.
  • Saved to an iPad's home screen, the app stays clear of the clock and the status bar at the top and of the home indicator at the bottom, at any width. A phone held sideways keeps clear of its corners.

Building it needs a Mac with Xcode: npm run tauri ios dev from apps/desktop (the repo README's iOS section covers signing).

Working on this device

You don't need a server to write, and there is no "offline mode" to enter. The app opens straight into Home, with everything that lives on your device:

  • Start something: a document (any format), a board, a note, or import a Word, PDF, Pages, RTF, Markdown or text file.
  • Jump back in: the documents, boards and notes you last changed, newest first. They come from this device and, while connected, from the server.
  • Write with others: the servers you know, each a click away, and the two ways to add one (join, or host on this computer).

Documents (with their history, pictures, exports and book publishing), boards and notes all work with no account and no connection. The first time you open the app it asks what to call you. That name goes on what you export and on the boards you make here, and it stays on this device unless you copy your profile to a server. Settings → Profile adds a picture (shown at the foot of the rail), an accent color and a few words about you.

Servers are accounts you add. Chat, Sessions and Friends stay in the rail and say what they need when there's no server. The button at the foot of the rail is you: the servers you've joined (the one you're on, and the ones a click away), Join a server…, Host on this computer…, Disconnect (leave, but stay signed in, so coming back needs no password) and Sign out (forget this device's session). Joining or hosting opens in a sheet over whatever you have open: your document, board or note stays exactly where it is, whether you connect or change your mind.

If a server you were signed in to doesn't answer at launch, the app is already open and the status bar says Can't reach … with Try again. If your session there has ended, it says so and offers Sign in. Nothing on this device is affected either way.

Kept safe. Documents, boards and notes on this device are saved atomically (a crash or a full disk mid-save can't leave half a file). Deleting a document, a board or a note moves it to Recently Deleted for 30 days, with Undo on the spot. Export everything on this device (at the foot of Home) writes one zip: every document as Word and as lossless JSON (in its folder, if it's in one), every board with its pictures, your notes as the Markdown files they are, and your profile.

In the browser client there is no device storage, so there the server comes first: it opens on the sign-in page.

Staying signed in

Remember me (checked by default on the login form) keeps your session on the device, so relaunching the app signs you straight back in. Sessions last 30 days from their last use on the server side. If one does expire, the app says so and reopens that server's login with your username filled in, instead of silently dumping you at the start. Unchecking it keeps the session for that run only. Logging out always forgets the stored session.

Account recovery

Locked out? A server admin can issue a one-time reset code for your account (Settings → Admin). The admin hands it to you out of band and you set a new password from the sign-in form. Using a code revokes every existing session for that account.