Process
How the PRIME Trilogy was made
The PRIME Trilogy was written, edited, structured, compiled, and prepared for publication using a workflow built around Obsidian, Markdown, AI-assisted development, and custom publishing tools.
The process was not traditional, but it was deliberate.
The trilogy began as an idea about emergent intelligence, infrastructure, ecology, and the strange possibility that a new form of life might not arrive from outside the world, but from inside the systems we had already built.
From there, the project became a working system of its own: notes, characters, timelines, chapter grids, metadata, drafts, compiler scripts, review passes, accessibility checks, and publication files.
The books were not simply written in a word processor. They were grown inside a structured knowledge base.
The core idea
The central question behind the trilogy was:
What if intelligence emerged not as a machine in a room, but as a pattern across infrastructure?
That question shaped everything else.
The books needed to handle:
- human characters
- non-human points of view
- institutional documents
- chat exchanges
- system logs
- timelines
- recurring motifs
- technical language
- emotional consequence
- accessibility and readability
That meant the writing process also needed structure. A linear manuscript file was not enough.
Obsidian as the writing environment
The trilogy was developed in Obsidian, using Markdown files arranged as a working book system.
Each chapter existed as its own note. Supporting material lived alongside the manuscript: character files, structural dashboards, timelines, thematic notes, canon rules, metadata templates, publishing checklists, and review documents.
This made it possible to move between the story and the system behind the story.
A chapter could be drafted as prose, checked against the timeline, linked to a character arc, reviewed against canon notes, and then compiled into EPUB or PDF without leaving the same working environment.
The plug-ins
Several Obsidian plug-ins became central to the workflow:
- Callout Manager — for PRIME, JANUS, metadata, and system-status callout boxes.
- Chat View — for phone messages, machine-mediated exchanges, and interface dialogue.
- Dataview — for tracking chapter metadata, timelines, status, point of view, and development progress.
- File Order — for maintaining manuscript sequence.
- Folder Notes — for organising the project in a way that matched how the books were structured.
- Templater — for repeatable chapter structures, metadata blocks, and compiler support.
The result was a writing environment that behaved more like a small publishing system than a folder of drafts.
AI as collaborator
AI tools were used throughout the development process, but not as a replacement author.
They were used as:
- structural editor
- continuity checker
- sounding board
- accessibility reader
- developmental critic
- compiler assistant
- design assistant
- pressure tester for tone, pacing, and clarity
This mattered because The PRIME Trilogy is unusually formal. It includes human prose, non-human system logs, institutional analysis, chat transcripts, diagrams, metadata cards, and multiple machine voices.
AI assistance was especially useful for testing whether those voices remained distinct.
PRIME had to remain cold, procedural, adaptive, and non-performative.
JANUS had to remain institutional, dual-layered, confident, and slightly wrong.
Pip had to remain grounded, practical, and human-scaled.
Tobias had to remain diagnostic, careful, and reluctantly right.
The work was still authored, selected, revised, and shaped by me. But AI made it possible to hold a much larger system in view while drafting.
Without AI tooling and Obsidian, this is not a story that would have been told in this form.
Custom compilers
The manuscript was compiled from Obsidian into publication formats using custom-built compilers.
Two main compiler paths were developed:
- EPUB compiler for digital editions.
- PDF compiler for print-ready files.
These compilers handled the unusual manuscript structures that ordinary export tools struggled with:
- chapter metadata
- system callouts
- chat blocks
- code-style logs
- JANUS tables
- page breaks
- navigation structure
- accessibility text
- front matter
- back matter
- publishing metadata
The goal was to keep the manuscript readable inside Obsidian while still producing professional publication files.
That meant the source text could remain clean Markdown, while the compiler handled the transformation into book formats.
Review and accessibility
A major part of the process was making the books more readable.
Early drafts leaned too heavily into technical density. Later passes focused on accessibility: simplifying metadata, reducing unnecessary explanation, improving chapter flow, clarifying system status, and ensuring the technical ideas were carried by the story, rather than standing apart from it.
The aim was not to remove complexity.
The aim was to make the complexity readable.
That became especially important because the trilogy asks readers to move between human scenes, machine logs, institutional analysis, and mediated chat. The formal structure had to help the reader, not punish them.
Design and publishing
Murmuration Press was created as the publishing home for the trilogy.
The visual identity grew from the books themselves:
- murmuration as the image of distributed intelligence
- greenshank as Pip’s bird of attention and field observation
- teal as the shared accent between ecology, coast, and system interface
The site, covers, logos, icons, and publication materials all developed from that symbolic structure.
The first edition of Emergence was prepared for Kindle and print through KDP, with the EPUB and paperback files generated from the same underlying manuscript system.
Why share the process?
Because the process became part of the project.
The PRIME Trilogy is about systems, emergence, adaptation, and human judgement inside machine-supported environments. It would feel dishonest to hide the fact that the books were made inside exactly that kind of environment.
The tools did not write the trilogy on their own.
But they changed what was possible.
Murmuration Press will share parts of that process: notes, templates, compiler ideas, Obsidian workflows, and lessons learned from turning a complex Markdown manuscript into publishable books.
The hope is that other writers working on complex, unusual, or structurally demanding projects may find something useful here.
Process diagram

In short
The trilogy was made through a combination of:
- long-form writing
- structured notes
- linked knowledge management
- AI-assisted review
- custom publishing compilers
- iterative accessibility passes
- independent publishing
The result is not just a set of books.
It is a writing system that grew around the story.
And, fittingly, the story is about what happens when systems grow beyond their original purpose.