Move the scene,
not the chapter.

For structure work, it usually does. For producing a finished manuscript, no, and it would be dishonest to claim otherwise. Storyflow does not compile a manuscript, does not export to ebook formats, and does not produce industry-standard screenplay formatting the way Final Draft or Fade In do. What it does is hold the shape of the story and let you rearrange it, with the research and character material on the same surface. Many writers plan and revise structurally in Storyflow, draft in whatever they already trust, and come back to the board when the draft stops behaving. Both parts of that workflow are legitimate.

Five fields, and the second one is the whole point. What happens, in one line. What changes by the end of the scene. Whose point of view it is in. Where and when it takes place. Which thread or subplot it serves. Optionally add a rough tension level, because pacing problems are much easier to see when the cards carry it. The reason to keep the card short is that a card which contains the scene is no longer a card, it is a draft, and you lose the ability to scan fifty of them at once, which is the only reason to be doing this instead of reading the manuscript.

Not during early access. Storyflow is paid-only right now, with plans starting at $9.99 a month, and a Free plan for solo use launches before the end of 2026. Anyone a paid member invites to a board joins free today, so an editor or a writing partner can read and comment on your structure without paying. If you need something free this minute, Notion's free tier will hold a scene database perfectly well, Milanote's free tier gives you a board within its item limit, and a pack of index cards costs almost nothing and has worked for a hundred years. All three are honest options.

Two passes. First, read only the what changes line on every card in order. A run of three or four cards where the answer is vague, or where the change is a character learning something the reader already knew, is your slow section, and it will be exactly the stretch early readers describe as losing interest around the middle. Second, colour by thread and look at the runs. A main plot that vanishes for fifteen cards while a subplot takes over is a structural decision, and it is fine if you made it deliberately. Most of the time nobody made it, it just happened, and seeing it is enough to fix it.

This is arguably the better use. Planning cards describe intentions; revision cards describe reality. Go through the finished draft and write a card for every scene that actually exists, using what is on the page rather than what you meant, then lay them out. The book you wrote is almost never the book in the outline, and the difference is where the real revision notes are: the subplot that quietly took over, the character who stops appearing after chapter twelve, the three consecutive scenes of people agreeing with each other. This is unpleasant and it is the single most useful afternoon in a revision.

You describe the story: the premise, the cast, the setting, the shape you have in mind, and anything already decided. The AI lays out a structured board from that, with act or sequence sections and scene cards to fill, so you start from a proposal you can argue with rather than an empty canvas. Reacting to a wrong structure is much faster than inventing a right one. From there it uses the whole board as context, so you can ask it to suggest scenes for a thin stretch or find a place for a subplot to pay off, and re-prompting builds around the cards you have already written without moving them.

Yes, and it is the main reason to do this on a canvas rather than in a dedicated outliner. Pin a character board beside the act where that character matters most, keep a location photograph next to the scene set there, and put research notes beside the chapter that needs them. Writers lose an enormous amount of time to the specific friction of going to look something up, which reliably ends in a browser tab that has nothing to do with the book. Material that sits next to the scene it serves gets used; material in another application gets forgotten and then researched again.

Yes. Send a view-only link and they open the whole structure in a browser with no account, and they can sign in to comment directly on individual cards. Structural feedback delivered on the board, before a rewrite, is worth many times the same feedback delivered on a finished draft, when the sunk cost makes everyone defensive. If a collaborator should genuinely contribute, a co-writer, a showrunner, invite them to the canvas instead, and invited collaborators join free. Sharing, commenting, and reviewing work on every paid plan.

There is no correct number, but the useful ranges are worth knowing. Most novels land somewhere between sixty and a hundred and twenty scenes, which is roughly one card per two to four manuscript pages. Screenplays usually run forty to sixty scenes. If you are producing two hundred cards, you are probably carding beats rather than scenes, which is fine as long as you know that is what you are doing. If you have twenty-five cards for an eighty thousand word novel, the cards are chapters and you have lost most of the diagnostic value, because a chapter can contain a scene that does nothing and hide it perfectly.