Reverie

My Thoughts on AI

2026-06-06T00:00:00+00:00

AI, what a subject. It draws a lot of emotive reaction. So I thought I’d sit down, think it through properly, and give you my views.

When I look at AI, there’s a part of me, the sci-fi part, that thinks it’s very cool, and that it’s an exciting time to be alive. And then there’s the thriller part, the one that thinks the dystopian robot takeover is inevitable. Exciting and scary at the same time.

But we’re probably thinking about it from the point of view of the author. And I do have strong views here. In creating Reverie, I wouldn’t have made such a big deal of the feel of the editor if I didn’t want people to write in it. It’s for us, the humans, to write. It’s certainly not a place for AI. If I wanted to build an AI novel writing system, it would look very different.

But I don’t want to build one, because I don’t think it has a place. I genuinely worry about where all this is going. AI is generating so much content now, and then learning from its own output, that I think it’ll become some self-feedback monster that homogenises everything into the same plain and boring prose. You can already see it in AI images. You can’t always say why, but you feel it. There’s something a bit too clean about them, a bit too predictable, and everything starts to look the same once everyone is using the same tools. Writing is going the same way. When a tool finishes your sentences for you, the voice that comes out isn’t quite yours. It’s yours mixed in with everything the machine has ever read. But that’s not even the point.

Writing is about telling a story we want to tell. It’s a craft. It’s something that wants to jump out of us onto the page. And AI generated stories are really just being made to shortcut at best, and at worst so people can ‘churn’ out content for pure financial gain.

So that said, I can promise you that Reverie is about you. The human. About helping you tell your story without an AI in the way to distract you. To help you get in the flow and tell that story.

Mark

Organising a Novel Without Outlining: A Discovery Writer's Approach

2026-06-02T00:00:00+00:00

Structure for the writer who finds the story by writing it.

Being a discovery writer works exactly the way you want it to, right up until it doesn’t. You sit down, you write the scene that’s in your head, you write the next one, and you keep going. No plan, no outline, no folders to fill in first. That’s the whole point. The story comes out because nothing made you stop and decide where it was supposed to go.
And then one day the draft is sixty thousand words long and you need to find the scene where she finds the letter. You know it’s in there. You scroll. You scroll back. You use find, but you can’t remember the exact words, so you search for “letter” and get forty hits. Or worse: you realise two scenes are in the wrong order, and fixing it means selecting three thousand words without losing a paragraph, cutting, scrolling, finding the seam, pasting, and reading the joins to make sure you didn’t break anything. The thing that made drafting easy is now making revision hard.
I wrote a whole post about why I left Scrivener. It’s a powerful, carefully made tool, and for writers who plan before they draft it’s a great one. I’m just not that writer. I’m a discovery writer, and being asked to organise before I’d written anything just screeches me to a halt. But near the end of that post I also wrote that there’s a real need for the organisational side of novel writing, and that it was a problem I’d love to take a proper run at someday. This is the first part of that run. The trick was to do it without recreating the thing that had stopped me: structure you have to build before you write.
The structure is already there</h2>
You didn’t plan your draft, but you didn’t write it as one shapeless block either. When one scene ended and another began, you marked it. Maybe you typed a scene break: three asterisks, the little divider that’s been signalling “time and place have shifted” in manuscripts for a century. Maybe you wrote a heading. Maybe a chapter title. You did this without thinking about it, because it’s how writing works: you put a small mark between the thing that ended and the thing that started.
That mark is structure. You made it while you were writing, not before. Reverie reads it back to you.
Open the sidebar and you get a list of your scenes, in order, each one labelled with its first line. Click one and you’re there. That’s it. You didn’t build the list. You didn’t drag anything into folders or fill in a synopsis card. The list is a reflection of what you already wrote, surfaced at the moment you need it and absent every other moment.
This is the whole difference. The planning-first approach gives you a structure to fill in before you’ve written a word, and asks you to know where things go before they exist. Reverie waits until you’ve written, then shows you what’s there. One way asks you to plan up front; the other reflects what you’ve already made. For a discovery writer that distinction is everything, because the planning is exactly the part that never worked.
You decide what counts as a scene, too. A scene break always does; that’s what it’s for. Headings are up to you: maybe your chapter titles are the unit you think in, maybe it’s the sections under them. You tick the ones that matter to how you see the draft, and the list reorganises to match. Reverie isn’t deciding your structure. It’s reading the one you made.
What it costs</h2>
I have to be straight about the tension here. Reverie is built on the idea that there should be nothing on the page but your words. No binder, no corkboard, nothing sitting beside the text. A sidebar full of scenes is exactly the kind of thing that idea is built against, and I’m not going to pretend otherwise.
So here’s the trade. The sidebar is a panel. When it’s open, it sits beside your text and nudges the page over to make room. That’s a real cost. It’s chrome, and Reverie’s whole argument is that chrome is what pulls you out. What it buys is that on the day you can’t find the letter scene, you find it in a second instead of a minute, and the minute is the one that ends the session.
The resolution I landed on is that the panel is off until you ask for it. The default is still a page and nothing else. The structure is computed quietly whether you’re looking at it or not, so it’s instant when you open the sidebar, but it doesn’t exist on screen until you reach for it with a keystroke. You get the page you came for while you’re drafting, and the map when you’re revising. Those are different jobs, and they happen at different times, so the tool can be two things without being two things at once.
It also means the feature does nothing until you’ve given it something to read. Write one long unbroken draft with no breaks and no headings and the sidebar stays empty, because there’s no structure to show until you’ve made some. The empty sidebar is correct. It waits until you’ve done the part only you can do, then shows you what’s there.
Moving a scene</h2>
Finding a scene is one half. Moving one is the other. The list of scenes is also a list you can reorder. Drag a scene to where it belongs and the words actually move, the whole span lifted and set down in the right place, the joins handled for you. That’s the part that turns “I know these two scenes are in the wrong order” from an afternoon of careful cutting into a single gesture. The way scenes are detected is the foundation it stands on. The same drag moves a whole chapter when a chapter is the thing in the wrong place, so the order you discovered becomes the order on the page.
The manuscript at the end</h2>
There’s one more part, and it’s the one that makes the rest worth doing. Finding your way around a draft is good. Sending it is the point.
When the writing is finished, Reverie takes the whole folder and exports it as a single manuscript in the format agents and editors expect. Times New Roman, double-spaced, a title page with your name and the word count, each chapter starting a new page, scene breaks marked the way they’ve been marked for a century. You set none of it up. You wrote in plain Markdown the entire time, on a page that asked nothing of you, and at the end you get a file that’s ready to submit.
So the whole path runs inside one app. You start on a blank page and find the story by writing it. The marks you left become a map when you need one. The scenes and chapters move when the order turns out wrong. And when it’s done, it leaves as a manuscript a publisher can open and read, with no planning at the start and no second tool at the end.
None of this changes when you write or how. You still sit down to a page that asks nothing of you. You still find the story by writing it. The organising waits, the way it’s supposed to, until there’s something to organise. Then it’s there, made out of the marks you left yourself, asking you to plan exactly nothing.
— Mark

The Best Writing App for Novelists: What I Found After Years of Searching

2026-05-25T00:00:00+00:00

What I liked, what I didn’t, and why none of them were quite right.


I’ve been writing on and off for twenty years and developing software for longer. I’ve tried most of the writing apps you’ve heard of and several you haven’t. Some of them are very good. None of them were right for me, and it took a long time to understand why.
This isn’t a features comparison. There are plenty of those. This is what it actually felt like to sit down and try to write in each one.

Microsoft Word</h2>
microsoft.com/microsoft-365</a> | From $99.99/yr, £84.99/yr, €99/yr (Microsoft 365 Personal)
Where most of us start. Where most of us stay longer than we should.
Word is built for documents, not for writing. There’s a difference. The toolbar alone has more options than I’ll use in a lifetime. Margins, headers, page numbers, track changes, comment bubbles. I’d open it to write a chapter and spend ten minutes adjusting the view before typing a word. Yes, Focus Mode exists. It hides the ribbon and gives you a cleaner view. But bolting a calm room onto the front of a factory doesn’t make it a writing tool.
The page feels clinical. White rectangle, black text, blinking cursor. No warmth, no personality. It’s paper on a screen, and not particularly good paper.
What it does well: track changes is genuinely useful when working with an editor, and the file format is the lingua franca of publishing. But for the act of writing? For sitting down with a blank page and trying to make words happen? It’s the wrong room.

Google Docs</h2>
docs.google.com</a> | Free
I had access to it, so I used it for a while. It’s convenient. No install, no file management, just a browser tab.
But I never got comfortable writing fiction in a browser. The page felt flat, more like a spreadsheet with better fonts than a place to do creative work. And I could never shake the sense that my writing lived on someone else’s computer, in someone else’s tab, one accidental close away from breaking my train of thought.

Scrivener</h2>
literatureandlatte.com</a> | $59.99/£59.99/€69.99 one-time
The one everyone recommends. The one I wanted to love.
Scrivener is powerful. The binder, the corkboard, the inspector, the ability to organise your manuscript into scenes and chapters and move them around. For writers who plan extensively, who outline before they draft, it’s built for exactly that.
My problem was that I’m not that writer. I’m a discovery writer. I find the story by writing it, not by planning it. I don’t know where a chapter belongs until I’ve written the chapters around it. Scrivener wanted me to organise first and write second, and that stopped me cold. I’d open it, see the empty folders and the structure waiting to be filled in, and close it. The app became another thing to manage instead of a place to write.
There’s a real need for the organisational side of novel writing, things like worldbuilding, character tracking, plot structure. Tools like Obsidian fill some of that gap, though none of them do it in a way that feels native to fiction. It’s a fascinating problem, and one I’d love to take a proper run at someday.
When I did get past the setup and actually wrote in Scrivener, the writing surface was fine. Clean enough. But “fine” is a low bar for the place where you spend the most important hours of your creative work.
No AI, no subscription. Those are real strengths. If you’re a plotter, if structure helps you think, Scrivener might be exactly what you need. It just wasn’t what I needed.

Ulysses</h2>
ulysses.app</a> | $5.99/£5.99/€5.99 per month, Mac and iOS only
Beautiful app. The best-looking writing experience on the Mac for a long time.
I loved the library. Everything in one place, organised by groups, searchable, synced across devices. The Markdown editor is clean and well-made. Publishing to WordPress directly from the app is clever. The writing experience is pleasant.
Two things pushed me away. The first is the subscription. We live in a world now where everything is a monthly payment, designed to feel cheap in the moment but adding up to far more in the long run. Open your bank statement and count the direct debits. It’s exhausting. And a writing app is the worst place for it. During a dry spell, the subscription made me feel guilty for not opening the app. During a productive stretch, I’d wonder whether I was writing because I wanted to or because I was trying to justify the cost. A creative tool shouldn’t carry that weight.
The second thing is harder to pin down. Ulysses is a very good container for writing. But the page itself, the moment of sitting down and typing, felt the same as every other app. Clean, minimal, static. The words went in and sat there. Nothing about the environment made me want to stay longer or come back sooner.

iA Writer</h2>
ia.net/writer</a> | $49.99/£49.99/€49.99 one-time (Mac), $29.99 (Windows)
The purest of the minimal editors. iA Writer strips everything away until there’s nothing left but text.
It’s opinionated in ways I respect. A small set of carefully chosen fonts. No formatting toolbar. Focus mode dims everything except the sentence you’re writing. The design is rigorous and the philosophy is clear: fewer distractions, better writing.
I have a lot of time for iA Writer. Their stance on AI, building Authorship to expose machine-written text instead of generating it, is the most thoughtful response any writing app has made.
But iA Writer is a Markdown editor, and it asks you to think that way. You’re writing in a syntax, not on a page. For developers and technical writers that’s natural. For a novelist who just wants to sit down and write a scene, it’s a layer of friction between you and the words. The writing should feel like writing, not like formatting.
The deeper issue is what kind of minimal it is. iA Writer’s minimalism is clinical. Everything has been removed, and you feel the absence. The page is stark, the cursor blinks, and you’re conscious of the emptiness in a way that puts you on edge rather than putting you at ease. Both iA Writer and Reverie are minimal. But there’s a difference between a room that’s been stripped bare and a room that’s so well considered you settle into it without thinking. One leaves you alert and aware of yourself. The other lets you relax. And when you’re relaxed, the words come easier. Not because of anything the app is doing. Because your guard is down.

What I actually wanted</h2>
After years of switching apps I could finally name the thing that was missing from all of them. Not a feature. A feeling.
Every app gave me a surface to write on. None of them made me want to stay there. On the hard days, the days when the blank page wins, every editor felt the same. Static, clinical, indifferent. The cursor blinked. I stared at it. I closed the app.
I wanted a page that met me halfway. Not with suggestions or AI or gamification. Something subtler. A page that felt alive. That responded to the act of writing in a way I couldn’t quite put my finger on but could feel immediately when it was gone.
I wanted to open my draft and feel like I was continuing, not starting. I wanted the app to know when the words were flowing and to quietly, invisibly, make the room a little warmer. I wanted to look up after twenty minutes and not know where the time went.
No app I tried did this. Not because they were bad. Because nobody was trying.
So I built Reverie</a>.
— Mark

Introducing Reverie

2026-05-03T00:00:00+00:00

A writing app for the writer who wants to write but isn’t writing.

I’m a developer by trade. I’ve been one for over twenty years. But I’ve also been a writer, on and off, for as long as I can remember. The kind of writer who has half-finished essays in folders, ideas in notebooks, the feeling of “I should be writing more” that never quite turns into writing more.
A few years ago I noticed something specific. I’d open Scrivener, see the binder, the corkboard, the inspector, the project structure, and close it. I’d renew Ulysses for another year and barely write in it. I’d open a Google Doc and feel nothing. The cursor blinking on a flat white expanse about as inviting as a spreadsheet.
The tools were excellent. They were not the problem. The problem was that every time I sat down to write, the interface was asking me to do something other than write. Plan a structure. Pick a folder. Set up a project. Choose between fifteen formatting options. Decide where this paragraph “belongs.”
And on the days I did get past that, when I actually started typing, something else would pull me out within minutes. A notification. A spell-check underline. A sudden urge to adjust margins. The cursor blinking on a clinical surface that broke the spell every time my eye landed on it.
I wanted a page that asked nothing of me except that I write on it. And once I started, kept me there.
So I built one.
What Reverie is</h2>
Reverie is a writing app where the page feels alive. The cursor glows softly. The scroll settles with weight. Formatting animates into place. The page warms when you’re in flow and cools when you pause, all below the threshold of conscious attention. You don’t notice these things directly. You notice that writing in Reverie is different from writing in anything else.
What you type is what you see. A heading looks like a heading. Bold looks like bold. There are no asterisks, no hashes, no syntax to learn or hide. Just text, rendered as text.
Your writing is saved as Markdown, the most common format in writing tools today. Open your files in any other app, on any other machine, in twenty years’ time. They’re yours. There’s no database, no proprietary format, no cloud account, no lock-in.
What Reverie is not</h2>
There’s no binder and no corkboard, no structure you have to build before you write. When the work gets longer, save your files to a folder and Reverie treats them as a manuscript. Switch between documents with a keystroke. Word count rolls up across everything. No setup, and nothing on the page but your words.
It’s not Ulysses. No subscription. Pay once, own it. No account required.
It’s not a Notes app. It’s built for chapters and long-form work, not lists and shopping reminders.
It has no AI. No plugins. No themes store. No collaboration.
These are choices, not omissions. Every “no” is something I actively decided not to build.
Why now</h2>
I’ve been making things feel right on screen since I started shipping games twenty-five years ago. Most of that work is invisible: the weight of a scroll, the way light sits on a surface, the gap between an animation that feels alive and one that feels like a tech demo. Games teach you that feel is engineering. Nobody calls it that, but it is.
Writing apps have never done this work. They give you a white rectangle, a blinking cursor, and nothing else. Reverie is what happens when you bring that attention to a page.
No investors, no co-founders, no roadmap committee. A craft project that became a product because enough early readers said “I would pay for this.”
The bet I’m making is that if five minutes in Reverie makes every other writing app feel dead, writers will stay. The page is the product. Everything else is in service of getting out of its way.
Coming soon</h2>
Reverie isn’t ready yet. When it is, I’ll announce it here and email everyone on the homepage</a> list.
If you write, or want to, I hope it gives you a page worth opening.
— Mark

The Neuroscience of the Page

2026-05-01T00:00:00+00:00

Why I built a writing app that knows when you’re in the zone.

You know the moment it breaks. You never notice flow arriving, but the end of it is sharp and unmistakable. A notification slides in. A word count updates. You glance at a toolbar you didn’t need and suddenly you’re aware of the cursor, the font, the room, yourself. The sentence you were about to write is gone. It was forming, and now it isn’t.
You sit for a while. You re-read what you’ve got. You write something, delete it, write it again. The thing that was carrying you forward ten seconds ago has stopped, and no amount of staring at the screen brings it back.
Every writer knows this. The state where the words come easily, where you look up and an hour has passed and there are pages you barely remember writing. It’s real and recognisable. So is the knowledge that it’s fragile. That the wrong interruption at the wrong moment doesn’t just pause it. It ends it.
What’s less well known is that neuroscience can explain exactly what’s happening. Researchers like Arne Dietrich have shown that flow isn’t your brain working harder. The regions responsible for self-monitoring and self-criticism, the voice that asks is this sentence any good?, go quiet. The parts that actually do the work get sharper. It’s a measurable neural configuration, and it has a measurable vulnerability. Even a small visual distraction in the wrong place can trip the mechanism that brings the critic back online.
I read that research and asked a question that seemed obvious but that nobody in the writing-app space had asked: what if the app was designed around protecting that state? A page that’s trying, quietly, to keep your inner critic asleep.

The blank page is the enemy. Not because it’s empty, but because of what it does to your brain. Research on goal pursuit has shown that people push harder as they approach a finish line, and they’re far more likely to keep going if they feel they’ve already started. A fresh document with a blinking cursor in the top-left corner says you have done nothing. It’s the worst possible emotional starting point for someone who is already struggling to begin.
So Reverie doesn’t give you a cold page. When you open yesterday’s draft, the page carries a trace of where you left off. Not a summary or a note to yourself, but a warmth. A sense that work has already happened here. You’re continuing, not starting. The difference is subtle, and it matters more than it should.
As you write, Reverie watches how you write, not what. Your typing rhythm, it turns out, is remarkably specific. Research published in Nature Scientific Reports has shown that the pattern of pauses between keystrokes closely tracks whether words are coming fluently or whether you’re searching. Not the speed. The pattern. Steady, rhythmic gaps mean the language is flowing. Erratic gaps mean it isn’t.
Reverie reads that rhythm and responds through the environment itself. When the words are coming, the page warms so slowly you’d never catch it happening. When you stop, it cools. The changes are calibrated against perceptual research. Slow enough, peripheral enough, and small enough that your conscious mind never registers them. Your emotional brain does. You feel supported without knowing why.
I borrowed something from game design, too. When you hit a milestone, the page sometimes responds with a quiet visual moment. Sometimes it doesn’t. That inconsistency is deliberate. Predictable rewards stop feeling like rewards. Unpredictable ones keep the brain’s reward system engaged. It’s the difference between a loyalty card and a slot machine, applied with the lightest possible touch.
The whole system runs at a fraction of what a game designer would call noticeable. There are no fireworks, no confetti, no screen shake. The test I use is simple: spend twenty minutes writing in Reverie, then open another editor. If the other editor feels dead, if something is missing that you can’t name, I’ve got it right.

A few things Reverie will never do.
There’s no AI. No generation, no suggestions, no rewriting. The app exists to support your relationship with the page, not to replace it. If you want to stare at a sentence for ten minutes until the right word arrives, that’s writing. I’m not going to short-circuit it.
There is a goals panel. It tracks your word count, your session time, your time in flow. But it never shows up on its own. You open it when you’re ready, not before. The writer should never feel watched while writing. That’s the core design constraint. The moment you become consciously aware of the feedback, it activates exactly the brain region I’m trying to keep quiet.
Your files are plain Markdown. No lock-in. You can use Reverie for your morning pages and keep your manuscript in Scrivener. I’m not trying to own your writing life. I’m trying to be the place where the words come easiest.

Whether the calibration is right, whether the thresholds land where the research says they should, is something only real writers using it daily will tell me.
Your writing environment should make the act of writing feel subtly, continuously, almost imperceptibly better. Not through features or AI, but through a page that responds to you in ways you never quite catch it doing.

Reverie

My Thoughts on AI

Organising a Novel Without Outlining: A Discovery Writer's Approach

The Best Writing App for Novelists: What I Found After Years of Searching

Introducing Reverie

Coming soon</h2> Reverie isn’t ready yet. When it is, I’ll announce it here and email everyone on the homepage</a> list.</p> If you write, or want to, I hope it gives you a page worth opening.</p> — Mark</p>

The Neuroscience of the Page

Coming soon</h2>
Reverie isn’t ready yet. When it is, I’ll announce it here and email everyone on the homepage</a> list.</p>
If you write, or want to, I hope it gives you a page worth opening.</p>
— Mark</p>