Long documents usually start in a notes app, and they should. It is the lowest-friction place to begin, the structure has not been decided yet, and opening a word processor to write two paragraphs is a commitment nobody wants to make at that stage.
The problem is that nothing tells you when to move. There is no error, no warning, no point at which the app declines. It just gets gradually worse at a job it was never for, and you notice about four days after you should have.
The four points where it breaks
Around 3,000 words: navigation. Below that you scroll. Above it you need a table of contents, an outline you can jump around in, and the ability to collapse sections to see the shape. A long note with no outline is a document you can only experience linearly, and you cannot think about the structure of something you can only read from the top.
Around 8,000 words: revision. This is where you need to move section four above section two, see both versions, and change your mind. Notes apps are built for appending, not for restructuring, and the operation you actually want — drag this chapter there — is either missing or is a cut-and-paste with a real risk of loss.
As soon as anyone else reads it: change tracking. Comments, suggestions, a record of what changed between the version they read and the one they are reading now. This is not a nice-to-have once a document is being reviewed; it is the entire mechanism of collaborative editing, and notes apps have none of it.
Whenever the output format matters. A submission with a required layout, a citation style, page numbers, a specific typeface. If the document has to look a certain way when it leaves, you are producing a formatted artefact, and that is a different category of software.
If you have hit none of these, stay where you are. Moving early costs more than it saves.
What a notes app is genuinely better at
Worth saying, because the answer is not "always use a word processor".
Everything before the first draft. Fragments, quotes, half-arguments, the paragraph you wrote in the wrong order. Research and thinking are exactly what a notes app is for, and doing them in a word processor produces premature formatting — you start choosing headings for a thing whose shape you do not yet know.
Anything with code in it. Word processors handle code badly. Smart quotes, autocorrect and automatic capitalisation are actively hostile to a shell command, and no amount of style configuration fully fixes it.
Documents that never leave. An internal note, a personal essay, a design rationale, a record for yourself. If nobody is reviewing it and nothing depends on the layout, the notes app is fine at 12,000 words and moving is busywork.
The notes about the document. Which stay behind when you move the prose out, and are the reason you should not delete the note when you migrate.
The split that works
Most people who write long things end up with the same two-place arrangement, and it is worth adopting deliberately rather than discovering.
The notes app holds the thinking. Research, sources, the outline, the arguments, the offcuts. It stays authoritative for that material for the whole life of the project — including after the document is finished, because it is the part you will want next time.
The document lives in a real document tool. Once the shape is decided, the prose moves and the notes stay. From then on the notes app is where you go to find the thing you read in March, and the document tool is where the writing happens.
The failure is trying to make one of them do both. Drafting in the notes app past the point it works produces a document you cannot restructure; researching in the word processor produces a folder of half-documents with no way to search across them.
Choosing the second tool
Briefly, since it depends entirely on the document.
Something with a real outline and restructuring. Scrivener is the long-standing answer on the Mac for book-length work, and the reason is precisely the second failure point above: it is built for moving sections around.
Something with comments and suggestions. If anyone reviews it, this is the deciding feature and it is why Google Docs and Word own that stage regardless of anything else.
Something with a stable file format, if the document will outlive the software. Markdown or LaTeX in plain files will be openable in twenty years. A proprietary format may not be, which is the same portability question that applies to your notes.
Something with citations, if you need them. A reference manager plus whatever it integrates with. Doing citations by hand in a long document is a decision you regret exactly once.
Two things worth doing before you move
Get the outline right first, in the notes app. Moving a document with a settled structure is a copy-and-paste. Moving one whose shape is still changing means you will restructure it twice, once in each tool.
Keep the note. Do not delete the draft when you move it out. It is the record of how the thing was made, it contains the paragraphs you cut and will want back, and it costs nothing to keep. Link it to the finished document so the two ends of the project are connected — a small use of linking that pays off years later when you write the follow-up.
The honest version
There is a strain of advice that says you should write everything in one tool, and it is wrong in both directions. Writing your research notes in a word processor is miserable. Writing a book in a notes app is a slow-motion structural problem.
Two tools with a clear boundary is not a failure of tooling. It is what having different jobs looks like.
Cyanote is the first half of that split and does not pretend to be the second. Notes nest into sub-pages, headings give a document structure, [[ links connect research to whatever it is for, and full-text search finds a quote you saved eighteen months ago. There is no outline pane, no change tracking, no comments, no citation manager and no page layout — so somewhere around a few thousand words of finished prose, the right move is to take the writing somewhere built for it and leave the thinking where it is.