The short answer: Publisher Rocket is the deeper desktop tool — keyword research with competition scoring, category research, and Amazon Ads keyword lists. KDSpy is a browser extension built for speed: open an Amazon bestseller page, click, get an instant read on the niche. If you research thoroughly once per book, Rocket. If you triage many categories quickly, KDSpy. Both are typically one-time purchases rather than subscriptions; check each vendor for current pricing, which we do not track. And both describe the shelf that already exists — which is the thing neither is built to see past.
Rocket's advantage is that it scores competition behind a keyword rather than just listing it, finds categories at a granularity Amazon's own browse tree hides, and exports keyword lists straight into an ads workflow. It rewards sitting down with it for an hour.
KDSpy's advantage is that it lives where you already are. You are looking at an Amazon category page anyway; one click turns it into estimated sales, review counts and rank spread. It rewards scanning twenty niches in the time Rocket would take on three.
Both read Amazon. Both describe supply: books that exist, keywords people already type, categories already populated. Neither carries any signal about demand that has not been met, because unmet demand has no book on the chart and no established search phrase. That is not a defect — it is outside the design of both — but it is the gap authors feel when the research is done and the question “so what do I actually write?” is still open.
Here is the distinction that matters, and it is not a slight on anyone's tool. Keyword and category research answers how do I get this book found. It assumes the book exists. Tropesmith answers a question that comes earlier and that no keyword tool attempts: which story do readers keep asking for and not getting. We count individual reader requests — a real person, in a review or a comment or a BookTok video, describing a book they want and cannot find — classify each one by trope, and score it against what the shelf actually supplies in that lane. Tropesmith reads a registry of 231,291 books, 2,679,139 Goodreads reviews, 67,499 book–trope tags across 11,575 fully tagged titles, and 721,971 classified reader requests, scored across 138 demand lanes — plus an hourly scan of 554 BookTok hashtags (72,641 videos logged, 27,618 reader requests read from BookTok).
What Tropesmith is not. It is not a keyword tool: it will not give you Amazon search volumes or seven backend keyword strings. It does not estimate anyone's sales or revenue. It does not build or bid your ads. If those are the jobs in front of you, the tool on this page does them and Tropesmith does not replace it. Plenty of authors run both, and that is the honest recommendation.
| The question you are asking | Publisher Rocket / KDSpy | Tropesmith |
|---|---|---|
| Keyword research with competition scoring | Rocket: yes · KDSpy: basic | No |
| Category research | Rocket: deep · KDSpy: quick | Partly — by lane |
| Estimated sales for a listed book | Rocket: yes · KDSpy: yes | No |
| Speed of a single look | Rocket: minutes · KDSpy: seconds | Minutes — it is a research read, not a triage click |
| Which story should I write next? | No | Yes — ranked by counted reader requests per lane |
| What are readers asking for and not getting? | No | Yes — 721,971 classified requests, with an underserved share per trope |
| How does my trope stack score against the shelf? | No | Yes — that is what a Map is |
| Is BookTok demand in the data? | No | Yes — 554 hashtags scanned hourly, 72,641 videos logged |
Tropesmith figures counted 2026-08-21. Rival capabilities are described from their published feature sets; we do not track anyone's pricing and deliberately quote none — check the vendor, and see our pricing page for ours.
The corpus behind every Tropesmith number on this page, counted on 21 August 2026.
Counted 2026-08-21 18:49 UTC by direct query against the Tropesmith corpus. Method: how we read reader demand · how a Map is built.
A Tropesmith Map scores your exact trope stack on live demand versus supply — and names the gap you can own.
Build my Map → See a real sample Map first → See pricing →Rocket if you do thorough research once per book and want keyword competition scoring and ads keyword lists. KDSpy if you scan many niches quickly and want a one-click read on an Amazon page you are already looking at. They overlap enough that most authors do not need both, and both are typically one-time purchases - check the vendors for current pricing.
On sales figures, both are estimating from Amazon rank, so both carry the same underlying uncertainty and neither should be treated as accounting. Rocket does more work on keyword competition, which is where the depth difference actually shows.
If you publish on Amazon and have not done keyword and category research, one of them saves real hours. Neither, however, answers what to write - they describe the shelf that exists, not the requests it has failed to satisfy.
Unmet demand. Both read Amazon's existing books, keywords and charts. A premise readers keep asking for and nobody has published leaves no trace there. Tropesmith counts those requests directly - 721,971 of them, classified by trope across 138 demand lanes.
That is the normal setup, and the one we would suggest. Tropesmith for the decision about what to write and which trope stack to build it on; Rocket or KDSpy for the keywords and categories once the book exists.
Every Tropesmith figure on this page is counted, dated 2026-08-21, and sourced from the reader-demand engine and book registry. Part of the market intelligence hub.