Free KDP metadata tool

KDP keywords for fiction: build seven phrases a reader might actually use

KDP keyword fields are not a place to repeat “book,” add an unrelated bestselling author, or paste the same phrase seven ways. Use this workbench to combine accurate story signals—trope, character, setting, tone, and reader promise—into seven candidates. Then validate them against Amazon's live suggestions and results before publishing.

Reviewed publishing tool · Updated July 2026 · No login · No invented search-volume claims

Build your seven keyword boxes

Choose only attributes that accurately describe the manuscript. The combinations update in your browser and are not sent anywhere.

  1. small town romance18 characters

What Amazon's current guidance actually says

Amazon's official KDP help page says keywords should accurately portray the book and reflect language customers use when searching. It recommends a logical word order, allows up to seven keywords or short phrases, and suggests useful dimensions such as setting, character type, character role, plot theme, and tone. Amazon's metadata guidelines separately recommend phrases of two or three words for best results. That is why this workbench keeps combinations concise instead of promising a secret list.

Amazon also tells publishers to test candidate terms in the store: inspect autocomplete suggestions and ask whether the returned books are relevant. This step matters because a phrase can sound plausible and still attract the wrong reader. The workbench cannot see private Amazon volume data, and it will not invent a number to make a phrase look authoritative.

Policy sources: verify the current rules on Amazon's keyword discoverability guide and book metadata guidelines immediately before publishing. WriteAIBook is not affiliated with Amazon or KDP.

Romance KDP keyword examples without the old matrix-page trap

“Romance keywords” is too broad to describe a specific manuscript, while eighty thin pages for eighty microgenres create little value. One useful page should teach the underlying job and let an author form phrases that match the actual book. For a closed-door grumpy/sunshine vineyard romance, seven candidates might look like this:

small town romance
grumpy sunshine romance
vineyard romance
competent heroine romance
closed door romance
workplace found family
emotional HEA romance

Do not copy that set onto an unrelated mafia romance, romantasy, or thriller. Change the attributes, search them in the relevant Amazon marketplace, and remove phrases whose results signal a different heat level, age group, ending, or subgenre. Accurate discovery is better than impressions from readers who immediately bounce.

Seven boxes, seven distinct discovery angles

1. Tone + lane + promise

Use the broadest phrase for the reader experience: heartwarming small-town romance, tactical LitRPG progression, or atmospheric folk horror. It should still exclude books that violate the central contract.

2. Trope + lane + setting

This is often the strongest fiction phrase because it mirrors how genre readers shop. Keep only tropes that generate repeated scenes, not a minor event mentioned once.

3. Character role + relationship engine

Character roles can surface a book for readers who seek a type of protagonist. Prefer concrete roles—hostage negotiator, older amateur sleuth, support-class crafter—over empty praise such as strong or amazing.

4. Setting + tone

A setting phrase earns a box when place changes the plot. “Stormbound mountain lodge suspense” is useful; “set in a town” is not. Avoid repeating a setting already prominent in the title unless another angle adds meaning.

5. Reader contract

Closed-door, low gore, fair clues, complete quest, and earned progression help qualified readers self-select. Only include the contract when the manuscript consistently honours it.

6–7. Cover important gaps

Use the final fields for another factual combination, not punctuation variants. Look for a character type, theme, or hybrid-genre signal not already carried by the title, categories, description, and first five phrases.

Terms to remove before you publish

Generic product words

Amazon specifically lists information common to most items in the category, such as “book,” among terms to avoid. A slot containing only “romance book” spends space without narrowing reader intent.

Quality and sales claims

“Best novel ever,” “bestseller,” “new,” and “on sale” are subjective or time-sensitive promotion rather than content metadata. The keyword fields are not ad copy.

Other authors and unauthorized brands

Do not use another author's name, a misleading comparison, or a brand you do not own. Relevance does not override metadata rules or intellectual-property responsibilities.

Amazon program names

Amazon's current avoid list includes Kindle Unlimited and KDP Select. Enrolment can be shown through the store's own program interface; it is not a keyword angle.

Spelling and punctuation variants

Do not spend multiple boxes on capitalization, plurals, punctuation, or spacing variants. Use the fields to cover different factual attributes instead.

Anything the book does not deliver

A popular trope that appears only in metadata attracts the wrong reader and risks misleading metadata. Every generated candidate must survive a manuscript check.

KDP fiction keyword questions

How many keyword phrases can I enter in KDP?

Amazon KDP currently provides up to seven keyword or short-phrase fields. Check the current KDP help page before publishing because platform interfaces and policies can change.

What are good KDP keywords for a romance novel?

Use accurate combinations of relationship trope, character type, setting, tone, and reader promise that are not already wasted on generic words. Examples include slow burn small town romance or grumpy sunshine vineyard romance when those phrases honestly describe the book.

Should I put Kindle Unlimited or KDP Select in a keyword box?

No. Amazon's current keyword guidance lists Amazon program names such as Kindle Unlimited and KDP Select among terms to avoid.

Does this tool provide Amazon search volume?

No. It organizes relevant phrase candidates. Validate them by checking Amazon search suggestions and the relevance of returned books before publishing; do not treat invented volume numbers as evidence.

Continue the publishing workflow