Reference

Every button, explained

Every button and menu line, with the words its own tooltip says. This page is read off the app, so it always matches the interface you have.

The sidebar

Where everything is sorted. A press opens its list in the column beside it.

Compress the sidebar and the list
Fold the sidebar to its icons (⌘\)
Search ⌘K: Search: a window, a note, a tag or a command (⌘K)
Home: Where to start, what to continue, what is due
Documents 3: Everything you added: the recent windows, your groups, your boxes
Recent: The windows you added, grouped, the latest first
What you marked to look at again: a window from its top bar, anything else from its menu
Groups: Groups of windows
Files kept aside in boxes, like folders: open them as windows when you decide, or never
Checkpoints: Where you stopped in each tree of windows
Ask the AI: your questions and their answers: the answer in words here, a page it makes on the workspace
View windows: Every question asked of the AI and the pages it made
The AI supervisors of your work: talk to them here
Workers: Workers hired to make pages, decks and courses from what your work holds, and their tasks
Coding agents that build and run tools on their own isolated machine: their windows, the tools they made and their runs
Outside services your agents can use: hosting, domains, email, payments, data — connected with their keys or by signing in
What your agents may build with: components, libraries, packages and services, kept from missions or added by hand
Every style and school on a real page: preview them, try representations and colours, make your own
Test each library the AI builds pages on: a tutorial and tests run where your pages run — and tell the AI what the results show
Every animation template your motion clips can use, with the library behind it: test each in a clip of its own and choose how much your clips use it
Test how a text sounds: one text, a voice, and one parameter varied across values, every variant spoken and kept to compare
Projects 1: Projects and everything that groups your work
Projects: Your projects
Research: Studies of one question
Experiments: Things tried before committing to them
An objective written in advance, and the orchestrators chosen for it, level by l: then its supervisors
Subjects: Directions followed while reading
Aims: Purposes and goals
Artifacts: Everything your work has made: written, generated, built, deployed and published, in one place and sorted your way
All: Everything your work has made, of every kind
Written: The documents you wrote and the diagrams you drew
Generated: The pages, decks and clips the AI made, and the texts it spoke
Built: The tools your agents built, kept in versions, and the results of their runs
Services: What runs and answers: the live apps and services your agents made
Deployed: Websites and apps put at a public address: on Orbital, a website host or a platform
Published: The pages you published on your blog, with where each came from and a way to unpublish it
Notes: What you wrote down, and your ideas
Notes: Every note, grouped by what it is on
Ideas: Ideas, sorted by kind
Highlights: Every highlighted passage, what was done with it, and where it is used
Tasks 2: What you have to do
Knowledge: Your vocabulary, references, people, places and things
Vocabulary: Terms and ideas you keep
References: Sources and where you cite them
People, places & things: Everything your work is about
Tags: Your tags: the tree, their groups, where groups cross, what is to do, and their map
All tags: Every tag, as a tree, grouped by kind
Groups: The groups your tags are sorted into, and their rules
Crossings: Groups crossed with each other, and what sits where they meet
To do: The places marked to do, by their state
Tag map: The map of your tags and what they hold
Your catalogs: a domain, a department, a client… — holding documents, terms, people, references and AI resources, and everything in your space sorted by type
All: Every catalog, as one tree
By type: Everything in your space, sorted by what it is: documents, vocabulary, people, places & things, references, tags, your work, notes, AI and files
Activity: Everything you have done in the app, day by day
Tools
Appearance: Light or dark, the text size, the contrast
Translate: Translate a text
Copied links: The links to windows you copied, ready to paste
Usage: What the data, the AI and the storage cost
Instant jump: Clicks land on a window at once, with no glide
Notifications: what the app told you · Right-click to make them silent
What your space runs for you: imports and AI generations — and the ones that failed
Trash: Deleted windows, which you can restore
Help: The guide, and what every button does (press ? then any button)
Settings: Your settings: the interface, the AI, the appearance
Your account: your details, sign out
Your account: your details, sign out
Complete: The interface level: how much the interface shows

The list and its filters

What you chose in the sidebar: its section's views as tabs to switch between, and the ways to narrow it.

Back to “Catalog”: where you were before
Next: nothing to go forward to
Fill the workspace by default: every window you go to grows to take the whole view
Compress the side column
Projects: Back to the list
Research: Studies of one question
Experiments: Things tried before committing to them
An objective written in advance, and the orchestrators chosen for it, level by l: then its supervisors
Subjects: Directions followed while reading
Aims: Purposes and goals
Filter: Filter: by kind, by date, by tag

The bar at the bottom of the workspace

Zoom, fit, the map, the layouts, and Add.

Previous window (Ctrl+Z) The water cycle, in one page
Next window (Ctrl+Shift+Z)
History: The windows you went to, the latest first
To re-examine: nothing marked yet. Mark a window from its top bar, anything else from its menu
Zoom out (⌥ + scroll)
Zoom in (⌥ + scroll)
Fit: Fit every window in view
Map: The map of the whole workspace
Layout: Arrange the windows: side by side, in a grid, one in focus
Full screen
Reload the page: anything not saved yet is saved first
Capture: draw a rectangle over the screen, and the picture is kept in your notes
Notes: your latest notes as a conversation, and a box to write the next one
Continue writing: write the next block of a Write window, right here (⌥B)
Appearance: light or dark, the text size, the contrast
Switch to Harbor
Ask the AI: a new question, in the AI view
Add: Add: a question, a document, a link, a note…
Fold this bar: keep it as one small button in the corner, so it hides less of your windows

A window's bar

Every window has one. At the Essential level it keeps a few buttons; the rest are in ⋯.

Ask the AI about this document: or about the words you selected in it
Pin this window (fast access in the right-bar window list)
A note on this window: and the notes already written about it
Add this window to something: a project, tags, tasks: and, at the fuller levels, everything else it can belong to — opened in its details; right-click for the menu
More actions for this window
Mark it to re-examine: it is listed in the dock and in Documents ▸ To re-examine
Fill the workspace: the window grows to take the whole view, and takes it again each time you come back to it
Full screen (Esc to leave)
Click to rename this window
Hide the bottom bar: for more room

Add to

One button, in every window's bar: a press opens the window's details at Organize; a right-click opens the same as this menu — projects, groups, tags, tasks.

File it in
Add to a project…: Put this window in one of your projects, or start a new one
Put in a catalog…: File it in catalogs of your own — a domain, a department, a client…
Add to a group…: Keep it with windows that belong together, to open them all at once
Add to a subject…: Follow one direction while you read: this window joins it
Source of a research…: Use it as a source of a research question
Put into an experiment…: Make it part of something you are trying out
Relate it
Tags…: Put keywords on it, to find it again with others
Relate to a term…: Link it to a word of your vocabulary
Relate to a person, place or thing…: Link it to a person, a place or a thing it is about
References…: The sources it cites, or the source it is
To do
Tasks on this window…: Make a task about it, or see the ones it has
Keep track
How important it is…: Rate it from one to five stars
Where I stand with it…: Mark it as to read, being read, or read

⋯ More

What a window can also do.

Note an idea…: Write down an idea about this window, filed with your ideas
Colour of this kind of window…: The colour every window of this kind wears
Dark pages: Read this PDF in dark mode — light words on a dark page; press again for white
Copy a link to this window: A link to this window, ready to paste in your notes
Bibliographic details…: Its title, authors, year and where it was published
Look up its reference…: Find its title, authors and year, and keep it in References
Replace with another version…: Put a newer or another version of this file in its place — its highlights, notes, tags and links stay
Stick mode…: Give windows fixed places: side by side, one above the other, in a grid
How this window works
Buttons in the bar
Delete this PDF (and its file)

Asking the AI

Under every question: what the answer should be, who it is for, how long, in which language — and, on the line with Ask, the model that answers and what the answer should cost.

Request templates: what the answer should be: a summary, a quiz, a timeline, a page…
What comes back: a text in this window, or a page beside it: the cheapest answer
For anyone: Who the answer is for: their level, how it is worded
How long the answer is: a ceiling, not a target
The language of the answer: none chosen: it comes in the question's language
Answer settings: Every setting of the answer, in tabs: content, look, audience, model, advanced
What this ask should cost, before it is sent: a text answer · 3.5k in · 900 out · An estimate: fewer than three answers of this kind to measure yet · Press for the breakdown
The model that answers: press to change it
Cancel: Clear the comment and question
Ask
Hide the options: keep only the box

Answer settings

Everything about the answer, in tabs.

Content
The answer: prose in this window
Length
Concreteness
Answer language
My answer preferences
Look
Design
Style gallery — every style as a real page
Customize the look
Libraries
Format
Signature
Audience
Who is it for? Audience & formulation
Who is answering? The speaker
Model & cost
The model that answers: opencode (local)
Service and context budget
Advanced
References: marked
Names: marked
Learned from my corrections
Playground

Keyboard

The keys that work everywhere.

Workspace
⌘K: the command line: open a window, tag, filter, create, export — type and Enter
drop a URL: anywhere on the workspace or the activity: it becomes an import
⌥ + Tab: cycle the recent windows (thumbnail strip) — ⇧ walks back
Esc: leave the window switcher without switching
⌥ + Q (hold): radial window picker around the mouse — release over one
⌥ (hold): radial picker of recent windows around the mouse
⌥ + Space: zoom out to the general view (⌥⇧Space zooms back in)
⌥ + arrows: pan the workspace, fluidly, for as long as they are held
⌥ + drag: marquee-select windows on empty workspace
⌥ + B: open / close the writing box — write the next block of a Write window from the workspace
Ctrl + Space: search the windows by name (blur + preview cards)
⌘Z / ⇧⌘Z: back / forward in the window history (pointer over the graph)
Right-click: on empty workspace: create windows, import a PDF or an EPUB…
?: this cheat sheet — it also lives in Settings ▸ Keyboard & gestures
Esc: close the search, a context menu or a picker (windows stay open)
Windows & selection
T: tag the selected text (the picker stays open for the next tag)
Right-click: on a selection: tag · comment · Ask the AI · dictionary
selection → Tag: link every selected window to a tag (Parent / Child descends the tree)
⌘/Ctrl + Enter: ask the question in the focused window
⇧⌥ + ↑/↓: bring the next/previous stuck window to the top level
← / →: in the picture viewer: the previous / next picture (Esc closes it)
Activity
Right-click: on a bubble: delete or detach the activity entry
↑ ↓: walk the entries in the activity filter — Enter focuses one
⌘B ⌘I ⌘K ⌘E: bold · italic · link · code, in the composer’s Aa mode
Tab / ⇧Tab: nest / un-nest a list item, in the composer’s Aa mode
PDF
Enter: jump to that page (the PDF page box)
Ctrl + wheel: the page scale inside the window — a pinch or ⌥ zooms the workspace instead
⌥ + drag: capture an area of the page as a picture (⬚ does the same without the key)
Writing
⌘B ⌘I ⌘U ⌘K ⌘E: bold · italic · underline · link · code, in a Write block
⌘⇧T: the type picker for the block the caret is in
⌘Z / ⇧⌘Z: undo / redo — the window’s own history, so a step survives a re-render
⇧ + Enter: end the line without leaving the block
Enter: on an empty list item: step out of the list
↑ ↓: in the writing box: from a row’s edge, walk the rows
Diagrams
⌘Z / ⇧⌘Z: undo / redo the text edit while typing in a block
Ctrl + B / Ctrl + N: undo / redo the diagram changes
Delete: remove the selected block(s)
Enter / Esc: new line in the block / leave the block
any letter: typed over the view: a text block is made where the mouse is
Gestures
right-click: the same first items everywhere: Tag · Open · Pin
drag a tag's port: onto another tag: puts that tag in this group (⌥ or ⇧: a relation)
⇧-click: in the activity with multi-select on: picks a range of bubbles
click outside: closes the tag picker, the palette, every popover
Undo (toast): a deleted window, note or tag comes back — eight seconds
⌥ + drag a tag node: moves that node alone — its members stay where they are

These pictures were taken from the app itself on 19 October 2026.

Orbital · Home · Sign in · Create a free account · Send us a message