Typographic graphic listing the History Co-Author front-end files with their line counts (framework 698, setup prompt 863, style pack 186, plus a version log) beside an example 13 by 13 Apollo 13 companion word search with a 12-word bank, no product screenshots

History Co-Author Review (2026): I Bought It and Read Every Kit File

Affiliate disclosure: BookPublisherTools earns a commission if you buy History Co-Author through the links in this review, at no extra cost to you. I bought the product with my own money, and I’ve tried to be as plain about the weak spots as the strong ones.

Last checked: 3 October 2026, during launch week.

How I reviewed it. I bought the front end and the History Genre Vault upsell and read every file in those downloads: all 18 setup steps, the writing rules, the style rules and all seven genre kits. I have not written a book with it yet. The plugins and the 29-module course weren’t in the downloads, and I haven’t used or watched them. So this review covers what you get and how it’s built, not how good the finished prose is.

The short verdict

History Co-Author is a set of long, carefully written prompt files for planning and drafting narrative history (true stories told like a novel) inside your own Claude or ChatGPT account. It is not software. What you’re paying for is structure: a planning process that makes you choose the story, angle, people and voice, and a writing method built to stop the AI from making things up about real people.

It’s a good fit if you:

  • want to publish narrative nonfiction on KDP and already pay for Claude or ChatGPT (or will)
  • like the idea of a step-by-step planning process with a recommendation at every step
  • will check the facts, quotes and sources yourself before you publish

Skip it if you:

  • write fiction (the same vendor sells separate fiction kits, and Sudowrite or NovelCrafter are built for novels)
  • want a button that outputs a finished, error-free book
  • don’t want a paid AI subscription on top of the price

At launch (October 2026) the front end costs $17, and the vendor says it rises to $27 after 7 October. There’s a 14-day refund window. For a nonfiction publisher who will do the fact-checking, that’s a low-risk price.

See History Co-Author’s current price and what’s included

What History Co-Author actually is

It’s made by Søren Jordansen (“The Copy Viking”), the vendor behind Cozy Co-Author and Steamy Co-Author. The sales page says it plainly: “You are not buying software.” You get three Markdown files to load into a Claude or ChatGPT project, plus two plugins (Claude Cowork and ChatGPT Work) and a written course. The genre is the kind of book that tells one siege, voyage or trial through the people who lived it, with notes and a bibliography at the back.

The files never name a particular AI model (“Nothing here is keyed to a particular model”, says the style pack), so the kit shouldn’t break as Claude and ChatGPT change.

What’s inside: the files I bought

The front-end download is a zip folder called Narrative History v1.0.1. Here is what’s in it:

FileSizeWhat it does
narrative_history_framework.md698 linesThe writing rules: genre, truth rules, voice, chapter process
narrative_history_setup_prompt.md863 linesThe 18-step planning conversation
narrative_history_style_pack.md186 linesBanned words and phrases, counts and limits
Version - Narrative History.docx1 pageVersion log: v1.0.1, “minor pre-launch fixes”

The line counts match the sales page (about 700, 860 and 186). Here’s what stood out.

Table of the seven History Co-Author genre kits with framework, setup prompt and style pack line counts and the number of narrator voices in each, counted from the v1.0.1 files

The setup prompt: 18 steps, every one a choice

The setup prompt runs one planning conversation from “give me your starting point” (a single word, 20 pages of notes, or “surprise me”) to five planning files. At every decision it gives three lettered options with a recommendation, one of which always sticks to your original idea, so most steps take a one-letter answer. The steps cover the story, whether the record can support it, the fresh angle, the story’s shape, whose eyes the reader follows, the voice and the title. Then it generates a Book Bible, a Research Bible, a People, Places and Entities file, a Chapter Outline and a Book State, one at a time, waiting for you to confirm each.

Research happens at three named moments, each with a hard cap: Step 3 (no more than a dozen source lookups), Step 9, the main pass (up to six helpers, 20 calls each), and Step 15 for gaps (up to three helpers). The AI is told to prefer transcribed original documents and to treat an encyclopedia as a route to a source, never the source.

The truth rules are the best part

This is where the kit earns its price. Every quotation in the Research Bible gets a label: EXACT, EXACT (QUOTED IN) (from a published history), PARAPHRASE, TRANSLATED or CANDIDATE (unconfirmed). Lines the research can’t trace go on a NOT TO BE USED list. While writing, the AI never searches; before each chapter it builds a one-page “Context Lock” with the exact quotes and figures that chapter uses, and you approve it first. Ordinary texture (weather, food, tools) can be written freely, but in the front-end kit nothing is invented about a named person: no made-up dialogue, no unsupported thoughts. (Some Vault kits work differently; see below.) Living people get stricter rules still.

None of that guarantees accuracy. AI models still get dates and quotes wrong, as the vendor’s own FAQ admits. But it’s a far better setup than “write me a history book about the Donner Party.”

Nine voices, labeled by author

The setup prompt has a Register Library of nine narrator voices, each labeled “similar to” a well-known narrative nonfiction author: The Close Witness (Laura Hillenbrand), The Reporter (Erik Larson), The Investigator (David Grann), The Elegist (Nathaniel Philbrick) and more. Each has an original sample passage about invented people, so you hear the voice before choosing. There are also three reading presets (the default plans 20 to 26 chapters of 3,000 to 3,500 words) and five story shapes, from the Approach to the Relay, with an explicit rule against falling back on plain chronology.

The style pack: a ban list for AI habits

The style pack bans words like “delve” and “tapestry”, clichés like “the die was cast” and “little did they know”, and em-dashes by default. It caps hedges like “must have” at one per chapter and has the AI count its own habits after each draft. That’s sensible editing advice, not a way to make a book undetectable (more on that below).

What wasn’t in my download

The plugins and the course aren’t in these zip files, so I can’t comment on them. The sales page says the plugins create your folders and run setup, chapters and close-out for you; without them you move the files yourself, which the instructions explain clearly.

The launch bonus: 50 book ideas

Order during launch week and you get a fast-action bonus: 50 book ideas with 252 collected sources. It showed up in my member area; the vendor says it’s tied to launch week, so don’t count on it later. The index PDF sorts the ideas into seven groups, from Exploration and Journeys to War and Survival, each with a brief, a date range and the authors in its source collection (the Donner Party, Apollo 13, Lincoln’s assassination, Waterloo, Salem and more). It points to a numbered source folder for each idea; I’ve only gone through the index so far.

How the workflow runs, start to finish

  1. Make a Claude or ChatGPT project and add the three files.
  2. Run setup: answer the 18 steps, approve the research passes, confirm the five planning files.
  3. Write one chapter per conversation (never more than four in one chat): approve the Context Lock, get the chapter and an updated Book State (capped at 150 lines), swap the old Book State out.
  4. Say “Close out the book” to get a Cast of Characters, A Note on Names, A Note on Sources, Chapter Notes and a Selected Bibliography.
  5. Fact-check, edit, format and publish.

The sales page says planning takes “a couple of hours” and writing “days, not months”. I haven’t timed it.

Pricing, order bump and OTOs (upsells)

History Co-Author launch funnel table: front end $17 at launch then $27, Ghostwriter License $27, Genre Vault $67 or $97, History Editor $47, Cover Creator $67 or $97, Promo Stacker $57, KDP Club $1 then $14.99 a month, plus a $20 a month AI plan

Prices at launch, October 2026, from the vendor’s sales, affiliate and upsell pages:

  • Front end, History Co-Author: $17 during launch week. The vendor’s affiliate page says the price rises to $27 after 23:59 EST on Wednesday 7 October 2026, and the product stays on sale after that. Expect $27 or more if you’re reading this later.
  • Order bump, History Ghostwriter License: $27. Lets you use the system to write for other people. Skip it unless you sell ghostwriting.
  • OTO1, History Genre Vault: $67, or $97 with the six genre plugins. I bought this one (details below).
  • OTO2, History Editor: $47. An editing framework for checking a finished manuscript. I haven’t seen it; consider it after book one, since continuity and notes are where AI-drafted books slip.
  • OTO3, Cover Creator: $67, or $97 with the A+ Author Toolkit. It builds prompts you paste into an AI image tool. Skip it if you already have a cover designer.
  • OTO4, Promo Stacker: $57. Book promotion training.
  • OTO5, KDP Club: $1 for the first month, then $14.99 a month (or $147 a year). A weekly live call. It’s recurring, so keep track of it.

The cost the price tag leaves out: you need a paid Claude or ChatGPT plan, about $20 a month according to the vendor’s FAQ. The front end and most of the upsell pages state a 14-day refund, and refund requests go to the vendor, not WarriorPlus.

Is the History Genre Vault worth it?

The Vault adds six more genre kits with the same file structure (framework, setup prompt, style pack, version log). They aren’t renamed copies; each has its own rules and voice library:

  • Ancient World: 15 voices, including registers similar to Mary Beard and Tom Holland
  • Battles and Campaigns: 15 voices, built “with the ground, the clock and the order of battle kept exact” (from its version log)
  • Christian History: 11 voices and a fourth preset, The Hourbook, for a single scriptural day or episode
  • Historical Biography: 14 voices, including The Lettered Life (similar to David McCullough)
  • Narrative Prehistory and Archaeology: 10 voices and two book shapes (vignette-and-evidence or reconstruction)
  • Young Readers History: 8 voices and a target-age setting (8–10, 10–12, 12–14 or 14–17) that changes sentence level, chapter length (about 1,200 to 3,000 words) and how dark the content gets

The truth rules differ between kits, and this matters. Historical Biography and Battles and Campaigns keep the front end’s rule: a named person’s thoughts and words come from their own record. Four kits loosen it on purpose:

  • Ancient World and Christian History let the AI fill in a named person’s thoughts, point of view, scenes and speech, as long as nothing contradicts the record (or scripture). They also let it create unnamed people.
  • Young Readers History fills in real people the same way, but nobody is made up.
  • Narrative Prehistory and Archaeology makes its people from the archaeological evidence.

Ancient World, Christian History and Prehistory tell the book to explain this once, in its preface or author’s note. If you want a strictly record-bound book, choose the kit with that in mind, or edit the filled passages yourself.

Buy it if one of those six is the book you actually want to write. If you only plan one general narrative history series, you don’t need it.

History Co-Author vs ChatGPT or Claude on their own

History Co-Author runs on ChatGPT or Claude, so the question is whether 1,750 lines of instructions beat your own prompts. For narrative history, I think they do:

  • Planning discipline. Most homemade prompts skip the angle, the story shape and the cast decision. This kit won’t let you.
  • A record system. Quote labels, a NOT TO BE USED list, locked name spellings and a Book State give the AI something to check against, chapter after chapter.
  • Back matter. Chapter notes and a bibliography built from the research trail, not invented at the end.

You could build a system like this yourself. It’s mostly careful rules and templates, and the kit’s value is that someone has already written and organized them. And the model underneath can still misread a source.

Other options, briefly. Sudowrite and NovelCrafter are subscription tools built mainly for fiction. Squibler has a nonfiction mode, but a 2026 write-up by Built&Written (a rival nonfiction tool) says it has no footnote or bibliography tools. For research alone, NotebookLM (answers from your own uploaded sources) plus Zotero (citations) is a solid setup. NotebookLM can also draft an outline or report from your sources if you ask. What neither gives you is a guided, step-by-step book setup with labelled-quote rules like this kit’s. Of the tools I looked at, History Co-Author is the one whose files are built around narrative history’s record-keeping: quote labels, locked name spellings and back matter drawn from the research trail. If another tool’s historical-book features matter to you, compare them on those points.

Claims to take with a pinch of salt

  • “No AI Slop. No Hallucinations.” The vendor’s own FAQ asks “Is that a promise of zero errors?” and answers “No.” The income disclaimer adds that AI models can get facts, dates, quotations and source references wrong. Check everything.
  • “#1 Amazon history series.” The screenshots show the books in Amazon’s Top 100 Free lists during a free promotion. That’s a real result, but it isn’t a paid bestseller rank. A Freebooksy listing confirms the first book was free on 19 September 2026.
  • Royalty screenshots. The income disclaimer says the results shown “are not typical”, and some testimonials come from the vendor’s other Co-Author products.
  • “Readers won’t know it’s AI.” I’d ignore this. Amazon KDP asks you to tell it when text is AI-generated, and its content guidelines say content created with an AI tool counts as AI-generated “even if you applied substantial edits afterwards”. The sales page has its own view on what that disclosure affects. Read Amazon’s rules yourself and answer honestly.
  • “$505 total value.” That’s the vendor’s own figure. The price you actually pay is $17 or $27.

Turn the same history topic into a companion puzzle book

This is where it ties into what we do at BookPublisherTools. A narrative history book leaves you with a ready-made word list. The People, Places and Entities file locks one spelling for every person, place and ship, and the close-out writes a Cast of Characters and Chapter Notes. That’s the hard part of a themed puzzle book, already gathered in one place. The files are AI-generated, though, so check every name and spelling against your underlying sources before you build puzzles from them.

With Puzzle Maker Pro, our Windows desktop puzzle generator, you can turn that material into a companion puzzle book:

  • Word searches from the cast and places. Paste the locked names into Standard Word Search as a word list, one puzzle per chapter or per group of people. The featured image above shows the idea with an Apollo 13 list (an illustration we drew for this article, not History Co-Author output). Start with getting started with Standard Word Search.
  • Clued puzzles from your chapter notes. Secret Word Criss Cross builds a real intersecting crossword with Across and Down clues from word lists with one Word: clue pair per line, for example AQUARIUS: lunar module the crew used as a lifeboat. The secret strip under the grid is optional (switch to Letter Hints if you don’t want one). Write the clues from your Research Bible so they match the book. Word lists for crossword variations explains the format.
  • Chapter-aligned sections. Puzzle Book Studio puts puzzles into sections with your titles and a solutions section. Creative builds a book from one puzzle collection. A book with several collections, such as one per chapter or word searches plus crosswords, needs Productivity, per the editions page. If you already own the Time Saver bundle or IPB Gold, you have Productivity-level Studio access already. A legacy per-module Time Saver licence gives that access only for its own module.
  • Kids’ editions. If you use the Young Readers kit, a matching activity book is a natural next step. See kids’ activity book interiors for KDP.

Turn your course or book into a puzzle book covers the whole process, and thematic word products from one theme shows how one list becomes several products. Every paid Puzzle Maker Pro edition lets you sell the books you make (commercial rights explained). Using words from your own book doesn’t settle the rights question on its own, though. KDP makes you responsible for copyright, trademark, privacy and publicity rights in everything you publish (KDP content guidelines), so give clues and word lists the same check as the book: quoted text, brand or product names, and anything about living people.

For step-by-step guides, see history crossword puzzle books, a history activity book for kids, history word search books that teach, Bible puzzle books for adults, WW2 and Civil War crossword books and how to write a history book with AI.

FAQ

Is History Co-Author software?

No. It’s three Markdown prompt files per genre, used inside your own Claude or ChatGPT project, plus optional plugins that run the files for you in Claude Cowork or ChatGPT Work.

Do I need ChatGPT Plus or Claude Pro?

Yes, a paid plan for one of them. The vendor puts it at about $20 a month.

What are the History Co-Author OTOs?

An order bump (Ghostwriter License, $27) and five upsells: History Genre Vault ($67 or $97), History Editor ($47), Cover Creator ($67 or $97), Promo Stacker ($57) and KDP Club ($1 for the first month, then $14.99 a month). All are optional.

Is History Co-Author worth it?

For a KDP publisher who wants to write narrative nonfiction and will check facts, yes, at $17 to $27. The planning process and truth rules are much stronger than anything I’d write myself in an afternoon. If you want hands-off books with no checking, no. KDP now limits new titles to two per book format each week (KDP Help: Create a Book), so a slow, well-researched book is a better use of a slot than several quick ones, and that’s the kind of book this kit is built for.

Does it guarantee accurate history?

No. It reduces invented details by keeping a labeled record and banning made-up quotes, but the AI can still make mistakes. Check quotes and figures against the sources.

What’s the refund policy?

14 days, full refund, requested from the vendor. The sales page says anything you made during that time is yours to keep.

Final verdict

History Co-Author is the most carefully built prompt kit for narrative history I’ve read. The truth rules, quote labels and close-out back matter are real work, and they’re in the files, not just on the sales page. Its limits: it runs on your own AI plan, it can’t stop the model making mistakes, and some of the marketing promises more than any prompt can deliver. If you publish nonfiction on KDP, it’s worth $17 at launch (or $27 after), with 14 days to change your mind. And when the book is done, the names and places in your research are ready for a companion puzzle book.

Get History Co-Author (affiliate link)

Shopping Cart