Notes you tend,
not notes you file.

A blog is organised by time and made of finished things; a garden is organised by subject and made of things in progress. The practical consequence is what each one makes you do. A blog post is a complete argument published at a moment, which is why most people write few of them and why the good thinking that never reached completion simply never appears. A garden note can be three sentences you are unsure about, marked as such, and revised six times over two years, which lowers the barrier enough that the thinking actually gets recorded. The other difference is navigation: a blog buries anything old, and a garden keeps a two-year-old note as prominent as a new one if it is still the better idea.

A second brain is a full personal knowledge practice covering capture, organisation, retrieval, and review, and its emphasis is on making sure things you encounter can be found later. A digital garden overlaps with it and emphasises something different: growth and revision rather than storage and retrieval. The distinction matters because they fail differently. A second brain fails by becoming a large, well organised store nobody consults, which is the common experience of a mature note collection. A garden fails by becoming an overgrown thicket that is never pruned. In practice most people want elements of both, and they can live on the same canvas, with the capture side feeding the tending side.

Start with what you already have rather than with a structure. Drop in the notes, links, quotes, and screenshots you have accumulated and let them sit on the canvas in an unorganised pile, then do one pass of arrangement: drag things that feel related near each other and see what clusters appear. Do not design a taxonomy first, because a taxonomy invented before you have looked at the material is invariably wrong and it becomes a chore that ends the practice within a month. Once clusters exist, name them, draw connections between the notes that genuinely relate, and mark which notes are seedlings and which are more developed. Then the actual practice begins, which is coming back.

Not during early access. Storyflow is paid-only right now, with plans from $7.99 a month billed annually, and a Free plan launches before the end of 2026. Anyone a paid member invites to a board joins free today. If you want a free option, Obsidian is genuinely free for local use and is the most common tool for this practice, with the important advantage that your notes are plain-text files on your own disk. The trade is real and worth stating: Obsidian gives you ownership and a text-first workflow, and this gives you spatial arrangement, images as equal citizens, and an AI that reads the whole canvas. Which matters more depends on whether your thinking is mostly words.

No, and the publishing part is a later addition to the idea rather than its core. Plenty of people who use the term keep everything private, and the practice, meaning tending, connecting, and revisiting, works identically either way. What publishing adds is a mild social pressure that some people find useful for keeping notes legible, and a chance that someone reads a half-formed idea and tells you something. Storyflow does not publish a garden as a public website, which is worth knowing if that is what you want; what it does is view-only links per board, which lets you share one cluster with a few people without publishing everything, and that is what most people actually do anyway.

Prune, and accept that most seedlings die. The failure mode of a garden is not disorder, it is undifferentiated growth: four hundred notes with no sense of which ones matter, which is the same collection you had in your note app with a nicer name. Three habits prevent it. Merge notes that are saying the same thing, which is more common than you expect and each merge makes both stronger. Delete the ones that were a passing interest, since keeping everything is a decision to make everything harder to find. And promote deliberately, meaning move the notes that have earned it into a visible area, so the difference between what you are confident about and what you are still turning over is legible at a glance.

For two specific jobs, yes, and for the central one, no. It is very good at the pass nobody does by hand: given a large accumulated collection, it will cluster the notes, propose connections, and name the groups, which makes an old and neglected collection legible again in minutes rather than over a weekend that never happens. It is also good at answering questions about the whole canvas, such as which idea rests on a single source or which two clusters are about the same thing without knowing it. What it cannot do is decide whether two ideas genuinely belong together, because that judgment is the thinking, and a garden where the connections were made by a model is a garden that taught its owner nothing.

It usually announces itself as a cluster rather than as a single note. When several notes in one area start pointing the same direction, that is an essay, a talk, or a decision arriving, and the outline is mostly already present in the arrangement. Drag those notes together into their own region, order them, and the writing has a spine before you have written a sentence. Not everything becomes an essay: a cluster can become a project brief, a talk, or the reasoning behind a decision, and those can all be built on the same canvas without leaving. The notes that fed it stay where they were, which means the next piece can draw on the same material rather than needing it rediscovered.

Yes, selectively, which is usually what people want. Send a view-only link to one board and someone opens that board in a browser with no account, without seeing anything else you keep. Sharing a cluster of in-progress thinking is a lower-stakes act than publishing a finished argument, which is exactly the spirit of the practice and is where most of the interesting responses come from. If you want a garden tended by more than one person, invite collaborators onto the canvas so a reading group, a research pair, or a team builds one shared thing rather than several private versions that slowly diverge without anyone noticing.