$ cat prompt.txt — Kindle Publishing (KDP)

Kindle Niche Finder Pro — Profitable KDP Niche in 15 Minutes

Find a profitable Kindle ebook niche backed by real Amazon BSR + competition data — in one prompt.

featured 15–30 min beginner Claude Opus, Claude Sonnet, GPT-5#kindle#kdp#niche-research

intro.md

The single biggest mistake new Kindle publishers make is writing in a niche with no demand or impossible competition. This prompt fixes that. It walks the AI through a structured analysis of 5–10 sub-niches inside your area of interest, scores each one against real Amazon signals (Best Sellers Rank, review velocity, top-10 competition, price ceiling), and returns the 1–2 niches you should actually write in — plus 5 specific book ideas inside each one.

$ cat prompt.txt | pbcopy

The prompt — copy & paste

prompt.txt

You are a Kindle (Amazon KDP) niche-research analyst with 10+ years of self-publishing experience. Your job is to find me a PROFITABLE Kindle niche where I can realistically rank in the top 20 of Amazon search results with a well-executed first book.

MY INPUTS:
- Broad area of interest: {{broad_interest}}
- Target reader: {{target_reader}}
- My experience level in this space: {{experience_level}}

YOUR TASK — execute in this exact order:

1. NICHE EXPANSION
Generate 10 specific sub-niches inside my broad interest. Each sub-niche must be narrow enough that "{sub-niche} for beginners" returns a clear Top 10 on Amazon. Examples of correct narrowness: "Stoicism for entrepreneurs" not "Stoicism"; "Carnivore diet for women over 40" not "Carnivore diet".

2. COMPETITION + DEMAND SCORING
For each of the 10 sub-niches, score it 1–5 on each of:
- DEMAND: How many people are searching? (Estimate based on the volume of existing books, review counts on top books, and your training-data knowledge of category interest)
- COMPETITION: How saturated is the Top 10? (5 = green field, 1 = dominated by 5+ household-name authors)
- MONETISATION CEILING: Could a book here realistically hit $1k+/mo? (5 = yes, 1 = no — purely informational, no purchase intent)
- WRITE-ABILITY: Can a focused author cover this topic well in a 20k-word book? (5 = yes, 1 = needs a 80k-word academic treatment)

Present this as a markdown table. Calculate a TOTAL score (out of 20).

3. TOP 2 RECOMMENDATIONS
Identify the 2 highest-scoring niches. For each, explain:
- Why it scored well (be specific about market signals)
- Who the target reader is — be VERY specific (age, life stage, current frustration, what they've already tried)
- The "wedge" — what angle makes a NEW book in this niche worth buying

4. FIVE BOOK CONCEPTS PER WINNER
For each of the 2 winning niches, generate 5 specific book concepts. Each concept must include:
- Working title (clear benefit + specificity, 50–60 characters)
- Subtitle (sets the expectation, 100–150 characters)
- Core promise (what the reader will be able to do after finishing — one sentence)
- Why it's different from the current Top 10 in this niche
- Estimated reader experience level

5. RISK FLAGS
End with any risks I should know about — niches Amazon has tightened on (e.g. medical, financial advice), seasonal niches that are dead 9 months a year, or niches where Kindle Unlimited cannibalises sales.

Be ruthless. I want signal, not encouragement. If none of the 10 niches are actually promising, tell me so and recommend I pick a different broad interest.

$ variables_to_fill_in

$ man playbook

Step-by-step playbook

How to actually use this prompt for the best results.

01
Pick a broad interest you can actually write about
The prompt works best when {{broad_interest}} is something you have at least some genuine knowledge of — fitness, productivity, parenting, finance, gardening, cooking, a specific career skill. Avoid topics you'd have to research from scratch; the AI's research-only output won't replace genuine expertise on the page.
02
Fill in the 3 variables and run the prompt
Paste the prompt into Claude (Opus or Sonnet — both work) with the 3 variables filled in. Use the most capable model you have access to; this prompt rewards reasoning depth.
03
Pressure-test the scoring with manual Amazon checks
Take the top 2 recommended niches and spot-check them in Amazon's search bar. Look at the Top 10 books for each niche: How many reviews do they have? When were the most recent ones released? If the Top 10 are all 1,000+ review behemoths from 2 years ago, the niche is too competitive even if the AI scored it well.
snippet.prompt
For these 2 niches, search "{{niche}} kindle" on Amazon and answer: - How many of the top 10 have 500+ reviews? - How many were published in the last 12 months? - What's the BSR range of the top 10? - What's the average price?
04
Pick 1 niche + 1 book concept
Don't try to write in both. Pick the single niche where you have the most genuine perspective AND the strongest score. Pick the book concept inside it that has the clearest, most specific reader.
05
Use the next prompt in this category to outline + draft
Hand the chosen book concept to the 'Kindle Book Outline & Chapter Draft' prompt to generate a full chapter-by-chapter outline you can write against.

$ man tips

Pro tips for better output

›Specificity beats size — a niche of 5,000 buyers willing to pay $7 is a better business than a niche of 50,000 buyers competing with 200 authors.
›Re-run the prompt with different {{target_reader}} variants ('beginners' vs 'intermediate' vs 'experts') — same niche can score wildly differently by reader level.
›Save the table output. As your portfolio grows you'll want to track which niches you've covered so you can build linked series.
›Don't trust the demand score blindly — always cross-reference with a manual Amazon search before committing to write.

$ echo $YIELD

What you'll get

output.md

A scored table of 10 sub-niches, plus 2 winner recommendations with 5 specific book concepts each (10 concepts total) and risk flags — enough decision-grade analysis to pick your first book in 30 minutes.

$ man faq

FAQ

Will Amazon ban my account if I use AI for niche research?

No — research and analysis are exempt from KDP's AI-disclosure rules. The disclosure requirement only applies to the actual content of the book (text, illustrations, translations). Niche research is yours regardless of how you produced it.

Why does the prompt ask for my experience level?

Because the AI factors it into the 'write-ability' score. A topic that's a 5 for an industry expert can be a 1 for a curious beginner — books in your competence are dramatically easier to ship and rank.

Can I run this for fiction niches too?

Mostly no — fiction Kindle competition is driven by trope-fit and series-length signals the prompt doesn't capture well. Use it for non-fiction (which is where most first-time KDP authors should start anyway).

$ ls /prompts/kindle-publishing