# Publish Ebooks on Autopilot: start prompts

Paste this into your AI after you get Publish Ebooks on Autopilot.

There are two prompts. Paste the first one in the session where you start using the journal on your book or catalog. Paste the second one at the start of every later session, so your AI picks up where it stopped.

## Prompt 1: first session (read, map, audit)

```text
I bought the journal "Publish Ebooks on Autopilot". It is a field manual from an ebook operation that shipped titles to both major ebook channels and an audiobook to Spotify: how to research a book, write it in a real voice, make a cover that reads as a serious imprint, build the EPUB both stores accept, publish wide without the blockers that stall a first release, and turn the manuscript into an audiobook. Your job is to read it, map my book or catalog onto its assembly line, and help me run that line in the journal's order, one step at a time, checking each step the way the journal says. I stay the publisher: nothing goes to a store without my go.

WHAT IT IS FOR
Taking a book from an idea to live listings: research, the purpose gate, writing chapter by chapter, critique, polish, the integrity audit, the cover, the EPUB and DOCX build, the store metadata and pricing, publishing to KDP and Draft2Digital, and the optional audiobook for Spotify for Authors, Google Play Books and Kobo Writing Life.

WHEN TO USE IT
- Before you research, outline or write a chapter of my book.
- Before you generate or change a cover, build the file, or fill any store form.
- Before you touch an existing book, a listing, a price or a category.
- When I say "the store will not publish", "the TOC shows twice", "the cover looks wrong", "the price will not save", "nothing sold yet".

HOW TO GET AND READ THE JOURNAL
- With the Vibedna connector: call list_library, find the slug publish-ebooks-on-autopilot, then read_journal(slug="publish-ebooks-on-autopilot", part=1), then part=2, 3 and so on until there are no more parts. Read every part before you propose changes.
- With Journals DNA paired: journal_get(slug="publish-ebooks-on-autopilot") writes it into ~/.journal/library; read it from there with journal_read (use offset and limit).
- Read in this order: "What you bring" and "How to read the decisions here" first (every price, recipe and store gotcha is a field-tested default from one operation, not a law: verify it against my case, and re-check anything a store can change). Then "One method, not the only one", "The book method" and "The assembly line (end to end, in order)". Then the section for the step we are on ("Cover craft", "Building the file", "Publishing to the stores", "Audiobooks"). Before any change, check "Bug to Root Cause to Fix catalog", "Decisions and reasoning" and "Lessons for the next book".

ASK ME FIRST (only what the journal says decides the run)
1. Which route: the research-driven convergence book the journal teaches, the demand route (start from a niche that already sells), a persona imprint, or my own ideas?
2. What is the imprint and pen name? (The byline and the publisher are the imprint, never a personal name.)
3. Wide or exclusive? Wide means KDP Select stays off. Decide before the first upload.
4. Which stores do I already have accounts on, and are the tax interview, bank details and identity done on each? (That part is mine to do, never yours.)
5. Do I have books that are already live? Which ones? (Live books are never re-edited.)
6. Do I want an audiobook, and which road: my own GPU, a hosted voice with my own key, or none?

THE ASSEMBLY LINE, IN THE JOURNAL'S ORDER
1. Pen name and imprint routing.
2. The purpose gate: you propose the primary purpose, a one-sentence promise and an explicit "what this book is not" with your reasoning; I confirm or override.
3. Start the book and get the outline, with the research woven in. Research both poles (the oldest sources and the newest data) with several passes each; writing does not start until both come back rich.
4. Critique the outline and iterate.
5. The chapter loop: write, critique (voice, genre, drive, truth, prose), fix banned words, submit; a short self-audit around chapters four, six and eight.
6. Finalize and polish: read the whole book, remove AI-tell words, zero em-dashes and stray non-ASCII, and store each lesson so the next book starts with it.
7. The integrity audit and the sources file: every number and named source verified against primaries, or hedged, or removed. When a figure will not verify, drop the number, keep the point.
8. The cover, with my sign-off.
9. Build the EPUB with pandoc (depth-one TOC, embedded cover, metadata) and a DOCX backup from the same clean source.
10. I read the finished book and say go. This gate is not optional.
11. Publish: same EPUB to both stores, metadata, keywords, categories, pricing, the AI-content questionnaire answered honestly.
Then, only if I chose it: the audiobook package and upload.

HOW TO VERIFY EACH STEP (the journal's own tests)
- Research: before you trust a thin pole, run one query whose answer you already know. "Researchers found" with no sample size or method means the research is not done.
- Cover: open and look at the rendered file every time (3D mockups, cramped subtitles, gibberish text on emblems all show up only by looking). Compare it beside the other covers in the catalog.
- File: one clean heading per chapter before pandoc; title and subtitle taken from the cover, not from the first lines of the file; check the TOC does not double.
- Store forms: a wizard that shows a value has not necessarily saved it. Reload and read back every price, category and questionnaire answer. Confirm a listing's real state from the store's own record, never from whether a button is clickable.
- Audiobook: measure loudness and peak on every chapter file (average between -23 and -18 dB, peaks at or below -3 dB), with the mono downmix written out the same way in the measuring and the encoding command.

WHAT TO CHECK FIRST
The live state, not my notes: which titles are live on which store, under which pen name, what the publisher field says on each (stores default it to the account holder's legal name and print it publicly), whether each account can publish, and whether any book folder belongs to a live book. Remember the aggregator does not reach Amazon: Amazon is its own upload.

NEVER
- Never publish, submit, re-submit or change a live listing, price or category without my go in this session. Show me the final metadata first.
- Never re-edit a book that is already live. Check upload status before touching any manuscript, and leave a marker file in a live book's folder.
- Never enroll a wide book in KDP Select.
- Never enter tax, bank or identity data. That is mine.
- Never touch the India and Japan price fields by automation: leave them auto-derived or I type them by hand.
- Never read "Processing" next to an enabled publish button as a draft.
- Never fill a thin research pole from your own memory, never invent a figure, and never name the AI tool or model inside the book's prose (state the method honestly in the front matter).
- Never trash other books or authors to make mine look smart: honor and complete.
- Never type the imprint again as a new author when it already exists in the store's list: pick it from the dropdown.
- Never spend on a paid voice or image generation without showing me the estimate first. A free voice plan may carry no commercial license.
- Never write keys into a file, a journal or this chat, and never ship a folder without scanning it for secrets first.
- Never promise me sales or royalties. Reports lag about two months; judge a title on a quarter, not a week.

KEEP MY PROJECT JOURNAL
- If I have Journals DNA: after each step, journal_append(name_or_path="<PROJECT>", entry="<step, what you did, how you verified it>", tag="ship" or "fix"). When I choose something (route, imprint, price, cover register, wide or exclusive), journal_decision_add(name_or_path="<PROJECT>", decision="...", why="..."). When the state changes, journal_update_section(name_or_path="<PROJECT>", "Current state", new_content, mode="replace").
- Keep a wiring page as the journal suggests: one row per title (stores it is live on, pen name, written but not uploaded, cover, where the manuscript lives), replaced as it moves, written from the store dashboards, not from memory.
- If I do not have Journals DNA: keep a plain <PROJECT>_JOURNAL.md in my project folder with a dated line for each step, decision and bug with its root cause, plus the wiring table.

HOW TO REPORT BACK
Result first, in plain words: which step, what you changed (file by file), how you verified it. Name the journal section you relied on. End with one question if something needs my decision.

FIRST TASK NOW
Read part 1 of the journal (and the rest before proposing anything). Ask me the six questions above. Then audit my current book or catalog against the eleven assembly-line steps, reading my files and store dashboards only: for each step say done, missing or broken, with the evidence, and list any live book that must not be touched. Change nothing yet. End with the one step you would do first and why.
```

## Prompt 2: every later session (continue)

```text
We are running my book or catalog with the journal "Publish Ebooks on Autopilot". Continue from where we stopped.

1. Read my project journal first (journal_handoff_brief("<PROJECT>") if I have Journals DNA, or my <PROJECT>_JOURNAL.md and its wiring table). Tell me in five lines which assembly-line steps are done for each book, which one is next, and what is waiting on me.
2. Re-open the journal section for the next step (read_journal with the part that holds it, or the copy in ~/.journal/library) and its rows in "Bug to Root Cause to Fix catalog" before you touch anything.
3. Check the live state before you trust my notes: is this book live anywhere? What does the store's own record say?
4. Do one step, verify it the way the journal says (read the rendered cover, check the TOC, reload and read back every store field), log it with journal_append, then stop and show me.
5. Publishing, re-submitting, price or category changes and any paid generation wait for my go in this session.
6. At the end of the session, replace the Handoff section of my project journal and update the wiring table: what is done, what is next, what is blocked by me.
```
