Worldbuilding a 4-Book Series Before Writing Book 2

I wrote the entire story of Etrath as a single novel. One book. The whole arc — from dying world to civilization to garden planet to utopia — crammed into one manuscript. It was too compressed. The world had too much to say for a single volume, and the pacing showed it.

So I made a structural decision: I took the first major arc of that original draft and turned it into a standalone novel. That became The Incubator. The rest of the story — the larger arc of the planet Etrath itself — became the framework for a four-book series.

That decision forced a second one. Before I wrote a word of Book 2, I needed to outline the full series. Not because I am a planner by nature — the opposite, actually — but because of how I had learned to write with AI.

Building planning-first AI systems, one pipeline at a time. Get weekly deep dives into local AI, Rust, and agent orchestration — no fluff, no hype.

How I Actually Wrote The Incubator

The Incubator took seventeen drafts. That number sounds painful, and it was, but the later drafts were short because of something I figured out partway through writing: the fastest way to produce a good story with AI assistance is a three-step loop.

Step one: write a skeleton. A scene-by-scene outline with the beats, the conflict, the outcome. Thin. No prose. Just what happens.

Step two: have the AI flesh it out. Feed the skeleton to an LLM and let it generate the prose. The output is readable. It is fun to read. It is also lifeless — technically competent, emotionally flat, written by a machine that understands story structure but does not understand stakes.

Step three: go back as an editor. Rewrite. Add the things the AI missed — the specific details that make a scene feel real, the dialogue that only works if you know the character, the moments of restraint where the AI would have over-explained. Cut the things the AI added — the filler paragraphs, the on-the-nose internal monologue, the overwritten descriptions. The goal is a story that reads like a human wrote it, because a human did write it. The AI did the heavy lifting of turning an outline into prose; the human did the actual writing.

This loop is what made The Incubator work. But it also taught me something about planning. The skeleton step only works if you know what the skeleton needs to contain. In a single novel, you can discover that as you go — write a skeleton for Act 1, see what emerges, skeletonize Act 2 based on what Act 1 established. The cost of a wrong turn is a draft revision. Manageable.

Why Discovery Writing Breaks Across Books

The single-novel approach works within one book. Across four books, it falls apart.

In the original single-novel draft, I could hold the entire story in my head. The constraints were self-evident because everything was in one file. When I split that into four books, I lost that. Book 2 needs to set up things that Book 3 pays off. Book 3 needs to remember things Book 2 established. Book 4 needs to resolve threads that span the full series.

If I discovery-write Book 2 the way I wrote The Incubator — skeleton, AI flesh, edit, repeat — I will introduce elements that feel good in the moment but create contradictions later. A character arc that peaks in Book 2 has nowhere to go in Book 3. A technology introduced in Book 2 needs to matter in Book 4, but I will not know that until I am writing Book 4, and by then Book 2 is published.

In a single novel, a wrong turn costs a draft. In a series, a wrong turn in Book 2 costs you Books 3 and 4 — or a retcon that breaks immersion for readers who were paying attention.

The seventeen-draft process for The Incubator was mostly structural revision — fixing problems that emerged from writing without a full plan. I did not want to repeat that process four times, each time constrained by what the previous book had already established.

What the Outline Actually Looks Like

The outline covers all four books — roughly seventy chapters total. Each chapter entry has:

  • Word count target — rough, keeps pacing honest.
  • POV character — the series has multiple POVs, and tracking who carries each chapter prevents the "wait, whose head are we in?" problem.
  • Plot thread tags — which ongoing threads this chapter advances, introduces, or resolves.
  • World state — where Etrath is in its planetary arc. A dying world, a rebuilding society, an active restoration, a stable utopia. This field forces every chapter to acknowledge where the planet is, even when the characters are not thinking about it.

The world state field was not in the first pass. I added it when I realized I could not tell from the chapter summaries alone whether the planet was still in crisis mode or had moved to the restoration phase. Without it, the series would feel like four books set in the same world rather than one story told at four scales.

The planetary arc, in brief:

  • Book 1 (The Incubator): Dying world. Ecosystem collapse, biological scarcity, genetic engineering as survival desperation.
  • Book 2 (The War of Ash): The transition. The immediate crisis is resolved, but the degradation continues. Society shifts from "how do we not die" to "what do we want to become."
  • Book 3 (The Gardeners): Active restoration. The survival technology becomes reconstruction technology. The conflict shifts to who decides what the restored world looks like.
  • Book 4 (The Prismatic Dawn): Utopia. The planet is healed. Stability creates its own problems — the old conflicts (power, identity, purpose) have nothing to hold them together.

Each book has its own plot and its own protagonist focus. But Etrath itself is the thread. A reader who finishes all four should feel the planet change the way a character changes across a single novel.

Why This Matters More With AI Than Without

Here is the part that surprised me. The AI-assisted writing loop makes planning more important, not less.

When I write with AI, the AI does not remember what it wrote three chapters ago. It does not track threads across books. It generates prose based on the context I give it — and if I do not give it the right context, it will confidently generate prose that contradicts something established earlier.

The outline solves this. Before I sit down to write a chapter, I feed the AI the relevant outline entries — the chapter summary, the active threads, the world state, the character arcs. The AI then has the full context it needs to generate prose that is consistent with the series. Without the outline, I would have to manually reconstruct that context every time, and I would inevitably forget something.

This is the same principle behind my planning systems in software: when the execution agent does not have persistent memory, the plan becomes the memory. The outline is not constraining my creativity. It is giving the AI enough context to be useful.

The Cost

The outline took three weeks. Three weeks of not writing prose. But those three weeks exposed structural problems that would normally appear in Draft 3 or Draft 4 of a manuscript. A character arc that goes cold for six chapters. A plot thread in Book 3 with no setup in Book 2. A thematic resolution in Book 4 that would feel unearned because the groundwork was never laid.

Finding those problems in an outline costs hours. Finding them in a manuscript costs weeks.

An outline is a cheap draft. A draft is an expensive outline.

What I Would Do Differently

I should have built the outline before writing Book 1. The Incubator is published now and its events constrain the rest of the series. Some constraints are productive — they give the outline boundaries. Some are not — there are things I would change about Book 1 if I could.

The other thing I would do differently is accept that the outline will change. I treated the first pass as if it needed to be right. It did not. It needed to be complete — every thread traced, every chapter accounted for — but "complete" and "correct" are different. The outline is a living document. Chapters will shift. Arcs will adjust. The point is that the adjustments happen in the outline, where they are cheap, not in the manuscript, where they are expensive.

The Principle

Planning and discovery are not opposites. They are stages. You plan the structure so that discovery can happen within it without breaking the structure. The prose, the dialogue, the specific way a scene unfolds — those are still found in the act of writing. The AI helps turn a skeleton into a draft; the human turns the draft into a story. But the skeleton needs to know where it is going, especially when the story spans four books and the AI has no memory of what happened in the last one.

Seventy chapters of outline is not seventy chapters of written book. It is the scaffolding that lets the written book stand — and it is the context that lets the AI write prose that fits inside a world, rather than prose that accidentally tears a hole in one.

The world these lessons come from lives in Etrath — a biopunk universe where biology is technology and harmony is power. The Incubator is available now.