Turning your D&D campaign into a novel, when what you have is a transcript, not a plot

A campaign is a transcript, not a plot

It is natural to look at a finished campaign and think the novel is mostly written: every line of dialogue, every decision and every reversal, preserved in order. It is not. What you have is a transcript — the minutes of the meeting — and a transcript is raw material the way a quarry is raw material for a cathedral.

The record is complete but it has no shape. Sessions are paced by the table, not by the story: the villain's reveal lands at eleven at night because that is when the table got there, the tensest negotiation is interrupted by a rules argument, and whole evenings pass in which the plot does not move because everyone was having too much fun shopping. None of that is a flaw in the campaign. All of it is a flaw in a manuscript.

The deeper structural problem is that events at the table happen for table reasons. The party went into the dungeon because it was Tuesday and that was the session. The rogue survived because a die came up nineteen. Fiction cannot run on that logic — in a novel, things have to happen because of who people are and what they want — so part of the adaptation is walking back through real events and building the causality the dice never needed.

Plan to cut about ninety percent

The proportion sounds brutal until you list what a session record actually contains: combat round by round, shopping, rules discussion, planning scenes where the party debates three approaches and takes a fourth, travel logistics, and the jokes. Somewhere inside a four-hour session is maybe twenty minutes of story that a reader who was not there could care about, and the book is made of those minutes.

The test for what survives is the one novelists already use: does this moment change what a character wants, what they know, or what it will cost them? A scene that changes none of the three is texture, and texture gets re-created in prose where the story needs it — it does not get carried over.

• Decisions that cost something, especially the ones the table argued about
• Relationships that changed: trust earned, trust broken, debts incurred
• Promises, threats and bargains — the load-bearing beams of any plot
• Reveals, and the earlier moments that planted them
• Consequences the party caused and then had to live with
• The two or three fights that were actually about something — kept as scenes, never as rounds

What the table never gave you: everything inside people

The biggest gap in even a perfect record is interiority. At the table, nobody narrates what their character privately fears, half-believes, or wants and will not say — the game happens out loud. Fiction is largely built out of exactly that inner material, which means the most important sentences in your novel appear nowhere in your source and have to be invented, scene by scene.

You will also invent everything that happened offscreen. Your villain existed only in the minutes the party shared a room with them; a novel's antagonist needs a plan that was running the whole time, and building it retroactively — so that every appearance in the record turns out to have been a move in it — is some of the most satisfying work in the whole adaptation.

And the physical world needs rebuilding. Rooms the DM sketched in one sentence, weather nobody mentioned, fatigue, smell, the passage of seasons: the table imagined all of it collectively and the record kept none of it. The reader gets only what is on the page.

Compress time, merge people, let the dice go

A long campaign accumulates a cast no novel can carry — every innkeeper, fence and quest-giver got a name at the table. Merge them. The innkeeper who passed on the rumour in session three and the fence who sold the map in session eleven can be one person with one agenda, and the story tightens every time you do it. The same goes for the calendar: eighteen months of Tuesdays can be one season in the book, and nobody will miss the four in-game weeks of travel.

The party is the harder version of the same call. You do not cut anyone's character, but a book has to choose whose story it is telling — one or two points of view, with the rest of the party seen from outside. An ensemble where six protagonists share every scene is what the campaign was; it is rarely what the novel should be.

Finally, dice outcomes become choices. A failed stealth roll is not a story event; a character whose impatience gets them seen is. Wherever the record says the dice decided something, the draft has to say a person did — which usually means promoting a random outcome into a flaw, a virtue, or an enemy's competence.

The workflow: from a pile of sessions to a first draft

Everything below assumes you can actually search the record. If the campaign lives in recordings, get them transcribed first — the transcription roundup compares the real options. And if it was transcribed elsewhere — a Craig export run through Whisper, another service's output saved as plain text, even a player's typed session notes — DM Scribe can import those transcripts on any plan, so the whole campaign becomes one searchable record.

1. Reread or skim the whole record once, in order, before you outline anything. Memory of a campaign is a highlight reel; the record is what happened. The gaps between the two are usually where the book is.
2. Build a beat list: one line per event that changed what somebody wanted, knew, or stood to lose, tagged with its session number. A whole campaign's beat list fits on a few pages — a thing you can hold in your head, which two hundred hours of audio is not.
3. Find the spine: whose story is this, what do they want, and what does it cost them? A campaign is everyone's story at once. A book has to choose.
4. Outline the book from the spine — the scenes the story needs, in the order it needs them — and only then map outline scenes back to sessions. Expect chapters that merge four sessions, and sessions that appear nowhere.
5. Draft every scene fresh, and never paste table dialogue. It reads like people playing a game, because it is. Write the scene the record describes in your own prose, and steal only the lines that were genuinely great — there will be a handful, and you already know which ones.
6. Keep the record open while you draft as a continuity reference: who knew what by when, what the baron actually agreed to, which names you have already used. Continuity errors are this genre's signature failure, precisely because the real events are checkable.

Before the rights questions: this is not legal advice

Stating that up front, the same way our recording-consent guide does, because the rest is more useful if you know what it is. We build transcription software; we are not lawyers or publishing professionals, policies and law change, and an actual publishing deal raises questions none of this touches. What follows is the shape of the two problems and the practical steps most writers take — not a ruling.

The two problems are genuinely separate. One is about the people at your table, and it is governed by fairness and friendship more than by any statute. The other is about Wizards of the Coast's intellectual property — and for that one there are current, public documents you can and should read yourself, quoted and linked in the table below.

Your players are collaborators — get it in writing

Their characters are their creative work. The dialogue you most want to steal was improvised by them; the decision the whole plot turned on came off the back of somebody else's choice. However much the novel adds — and the invention will be most of the book — a story built on a campaign is built on other people's contributions, and treating that as a courtesy problem rather than fine print is what keeps it from ever becoming fine print.

The practice that works: tell the table what you want to do before you draft, and get written agreement before you publish — written meaning anything you could point to later, a group email everyone answers included. Doing it early is the kindness. Doing it in writing is what protects the friendships if the book ever earns money or attention, because a generous verbal yes is exactly the thing people remember differently three years on.

• What is being used: their character, the character's name and history, and any specific scenes or lines
• Whether they want the character renamed or reworked — offer it, rather than waiting to be asked
• Whether they read the draft before anyone else, and whether that read carries a veto
• What happens if there is ever money — one plain sentence now beats a percentage argument later
• A real way to say no. If someone uses it, their character is written out or changed beyond recognition, without negotiation.

Wizards' IP versus your homebrew

Start with what is unambiguously yours: the world you invented, your NPCs, the plot, the prose you are about to write, and — with their agreement — your players' characters. A novel of a homebrew campaign is overwhelmingly made of those. What is not yours is Dungeons and Dragons itself, and two Wizards documents govern most of what fans can do with it. Both are quoted, with sources and dates, in the table below.

The Fan Content Policy is the generous one, and it is how campaign write-ups, fan art and streams exist: it lets fans build on Wizards' IP — but its own text requires that fan content be free, and it says plainly that selling Wizards-related content needs Wizards' permission. A novel you intend to sell is outside the Fan Content Policy by definition, whatever a well-meaning forum thread told you.

The 5.1 System Reference Document points the other way: Wizards released its full contents under a Creative Commons licence that permits commercial use, permanently, with attribution. But read what the SRD is before leaning on it. It is a rules reference — a licensed subset of the game's mechanics and names, not the whole of D&D — so if a specific creature, spell or place name matters to your book, check that it actually appears in the document rather than assuming it does.

Which is why the standard route for campaign-to-novel writers is the oldest one: set the book in your own world and rename anything that is not yours. It costs less than it sounds — the plot, the relationships and the characters carry the story, and none of them belong to Wizards. If you want the book to stay recognisably D&D and be sold, that is precisely the case for reading the current policy pages yourself and talking to a publishing professional, which this page is not.

The honest ending: some campaigns should stay campaigns

A campaign that was wonderful because of the people in the room does not always survive the people being removed. If the best nights were the bits, the chaos and the jokes — the things that only work live — the honest adaptation is not a novel. It is a campaign journal for the six people who were there, an actual-play edit if the fun is in the voices themselves — what publishing actual play involves — or simply a good recap archive the table can relive.

And be clear-eyed about the tools. If your campaign lives in a shoebox of paper notes, transcription software has nothing to offer you — the beat list is still the move, and a notebook holds one fine. Most campaigns make better memories than manuscripts. The ones that make both are the ones where somebody kept the record.