7 novels. 5 genres. 3 pen names. 9 months.

Status: Draft v2. This is the SOURCE doc. The X pinned thread, LinkedIn long-form, Product Hunt maker comment, 20BooksTo50K post, and paid-ad creative all derive from this. Edit this file first; everything else inherits.

Author: Mark Price (founder voice, first person)

Audience: Indie authors with outlines they never finished. Tech founders curious about long-context LLM engineering. Anyone who has ever stared at a draft folder full of "someday."

---

TL;DR (180 words, the version that goes everywhere)

I had ten novel outlines sitting in a drawer for years.

The books were already done. I knew the characters. I knew the world. I knew every climax. They lived in my head fully formed. I just couldn't get them out, because writing the prose took twelve months per book. So I quit writing.

Then I built the engine I needed. Nine months later, the books have come out. Seven of them, across five genres, under three pen names. Sword and sorcery (as Garrick Hale): two published, third in the queue. Paranormal romance (as JM Price): The Last Human. Sci-fi adventure horror (as Mark Wahlbeck): Earth Forgotten. Plus a sci-fi survival horror ready to publish, a YA academy fantasy I drafted in a single day, and a LitRPG I'm finishing in public over the next two weeks.

This is StoryForge. It launches today.

The strange part: when I read my finished books now, it feels like reading someone else's. Even though I know everything in them. The voice is mine. The story is mine. I just didn't have to type every word for a year.

That's the magic. That's the whole thing.

---

The long-form narrative (the source for thread + LinkedIn + 20Books post)

The pain

I wrote books before. I hated it.

Not the part where you imagine the world. Not the part where you outline the climax. Not the part where you figure out which trope you're subverting and which you're delivering on. Those parts I love. Those parts are why I write.

I hated the prose. Twelve months of putting one word after another to execute a thing I'd already finished in my head. Weeks of editing back-and-forth in Word docs. Then again for proofread. Then again. The book that lived in my head was sharp and fast. The path to getting it on Amazon was twelve months of grinding through prose that already existed in finished form. Just not on the page.

After a few books I quit. I had ten more outlines done. Full character arcs. Full plot structures. They sat. For years. The bottleneck was never the imagination. It was the keyboard time.

The thing I want most as a writer is to read my own finished book. Not to type it. To watch the characters do what I imagined them doing. The actual writing was always the price I paid to get there.

The build

I'm a coder. When AI got good enough that long-context models could hold a story bible across 80,000 words, I knew the engine I wanted was finally buildable. So I built it.

Nine months heads-down. The hard parts:

• Consistent voice across every chapter. Every other AI writing tool drifts. By chapter twelve you can feel the prose flatten into ChatGPT-default. I built per-chapter style validation that runs against your writing voice. The book sounds like you from page one to climax.
• Story bible tracking. Characters, settings, magic systems, who-knows-what-when. Every chapter validates against the bible. No more "wait, how does Marcus know this in chapter 8?"
• Outline-to-prose pipeline. Not "write me a chapter." Outline collaboration first, designed for commercial success. Hits the tropes the audience expects. Hits the plot beats that make the genre work. Ships KDP-compliant on submit.
• Two modes. Define every scene and every character if you want surgical control. Or rapid-release if you want to ship.

The receipts

In the nine months I was building, I dogfooded it on my own outlines. Across three pen names, five genres, seven books:

As Garrick Hale (sword and sorcery):

• The Wailing Spire. published on Amazon
• The Bound Fang. published on Amazon
• The Choir Abomination. drafted, in the publishing queue

Three books. Same characters across the trilogy. The same writing voice held from book one chapter one to book three's climax. (Book one is on KU and getting page reads, even though I've done zero marketing on it. The numbers are small but they are real.)

As JM Price (paranormal romance):

• The Last Human. published on Amazon

A different pen name for a different audience. Different voice spec, different trope contract, different reader expectations. The engine handled the genre switch without me having to retrain anything.

As Mark Wahlbeck (sci-fi adventure horror, my personal pen name):

• Earth Forgotten. published on Amazon

Drafted, ready to publish:

• The Eidolon Drift. sci-fi survival horror. 100,000 words across 20 chapters. The outline alone is detailed enough to be its own document; the engine took it from outline collaboration to KDP-compliant prose.
• A YA academy fantasy. Drafted in a single day. Not "wrote." Drafted in my voice, KDP-compliant on submit. One day.

Shipping in public right now (as JD Crane, my new LitRPG pen name):

• The Red Mark. LitRPG / progression fantasy / dark fantasy. 50 chapters across five acts, system mechanics, blood economy, full party arc. A few chapters left to close out. I'm finishing it in public over the next two weeks. If you've ever wanted to watch the engine ship a real KU LitRPG launch from outline to upload, this is the one.

The honest part

I have not marketed any of these books. Zero ads. Zero newsletter. Zero review-request campaigns. One book is on KU and getting page reads, the rest don't have meaningful traffic, and that's fine.

Critics will read that and say "see, AI books don't sell." Wrong question. The question this experiment was designed to answer is: does the engine produce finished novels in your writing voice that hold together as real books and survive KDP review?

Seven times, across five genres, three pen names, the answer is yes.

The "do they sell on KU" experiment is the next one. I'm running it on book eight, which I'm finishing in public over the next two weeks. The Red Mark, a LitRPG. If you want to watch a real KU launch from a few-chapters-left state to upload day, the build log is on X.

What's next

The Red Mark, in public. Over the next two weeks I'm finishing my LitRPG (50 chapters, the last few are queued) and uploading it to KU on launch day. I'll post the build log every step: which scenes the engine got right on first pass, which scenes I rewrote, the cost breakdown, the launch-day numbers. If you want to see what the engine does on a genre that punishes inconsistency more than any other, watch this one.

LitRPG-leaning features next. Stat blocks, level-up beats, system mechanics, progression curves. The genre is exploding on KU and the structural demands are exactly what the engine's bible-tracking and validation gates are built for. Beta testers in LitRPG / progression fantasy specifically, DM me.

Beta tester slots open. I have two beta testers right now (one fully dogfooding, one a few chapters in, both reporting positive). I want a small wave more before the next round of feature work. If you ship for KU, especially in romance, LitRPG, sword & sorcery, or sci-fi horror, the trial gives you 30 days to find out if this works for you.

StoryForge is live. 30-day free trial. Card on file for anti-abuse, never used until day 31. $49/mo after that.

---