A romance reader posted on Reddit a few months ago asking for what felt, to her, like a small thing.
“All I want is a simple… Is there rape in this book? Yes/No. Surely I’m not the only one.”
She wasn’t. The thread filled with replies from other readers — survivors, parents, mood readers, people just trying to enjoy a book without being ambushed by content they couldn’t handle that day. They were all asking variations of the same question, and they were all running into the same wall: nobody is systematically answering it.
We analyzed 2,122 reader comments across 30 Reddit threads — fantasy, sci-fi, romance, parenting, librarian, and bookseller communities — and the romance threads stood out for one reason. Romance readers haven’t just identified the problem. They’ve already built a vocabulary to solve it.
Tropes. Spice levels. Open-door versus closed-door. Slow burn. Forced proximity. Found family. Enemies-to-lovers with no third-act breakup. They know exactly what they want. They can describe it with surgical precision.
The discovery tools serving them are still stuck in 2010.
The DNF Problem
Romance readers DNF — “did not finish” — books at higher rates than almost any other genre community. Not because they’re picky. Because the genre has more tonal range than almost any other, and the marketing tells you almost nothing.
Two books with similar covers, similar blurbs, similar Amazon categories can deliver wildly different experiences:
- One is a closed-door regency with a slow-burn epistolary courtship
- The other is a five-pepper dark romance with on-page assault and a morally compromised hero
Both are shelved as “Romance.” Both might be excellent at what they do. But a reader who picks up one expecting the other isn’t mildly disappointed — they’re potentially harmed. And the only feedback loop is a one-star review written in frustration, which doesn’t actually help the next reader avoid the same surprise.
From the threads we analyzed: readers DNF books because of unexpected pregnancy plotlines they weren’t prepared for. Cheating that wasn’t signaled in the blurb. Death of a child. Detailed assault scenes in books marketed as steamy contemporary. Animal harm. Stalking framed as romantic.
None of these are objectionable in themselves. What’s objectionable is being blindsided by them.
The Romance.io Phenomenon
Romance.io is the closest thing to a solution that exists. Among the romance threads we analyzed, it was mentioned 118 times — more than any other tool except Goodreads itself. The praise was specific:
“Romance.io \u2014 you can search multiple kinks and tropes… honestly the best is the recs here. Get freaky with us!”
That comment got 84 upvotes. Just below it, a comment that probably tells you everything about romance.io’s product-market fit:
“If there was an app, I would use it instead of Goodreads!”
Sixty-six upvotes. From romance readers who actively want to switch — but the existing tools don’t serve them. Romance.io’s founder is even active in the threads, asking readers what tags to add. It’s the most engaged platform in the discovery space.
But romance.io has one structural problem the founder has acknowledged. From a top-voted reply:
“There are a few flaws with romance.io, mainly that everything is user added. This can mean if a book is less popular there may not be many things listed for tropes or triggers.”
That’s the collective action problem at the heart of every community-sourced metadata system. Popular books get tagged richly because lots of readers tag them. Less popular books — newer indie releases, midlist titles, backlist gems — get sparse tags or none. The reader who’s already off the bestseller list is on their own again.
Where Goodreads Fails Romance Specifically
Romance readers have unique reasons to find Goodreads frustrating. Across the threads we read:
- The recommendation engine is broken. One reader recounted being recommended The Very Hungry Caterpillar as “readers also enjoyed” on a high fantasy novel. The algorithm sorts by Amazon sales velocity, not by whether the book matches what you actually liked.
- Categories are gamed. Authors self-select into popular categories to surface in search results. “Erotic Romance” bestseller lists routinely include closed-door novels. “Clean & Wholesome” sometimes contains explicit content. The labels don’t mean what they say.
- Trigger warnings don’t exist. Goodreads doesn’t have content warnings as a structural feature. Some reviewers add them in their reviews — but you have to read the reviews to find them, and you have to read enough reviews to know they’re reliable.
- Mood and pacing aren’t searchable. Romance readers often want a specific emotional experience: “cozy and low-stakes,” “angsty with a happy ending,” “dark but cathartic.” None of that is filterable on the largest book platform in the world.
The MPAA Problem
Across romance threads, one phrase appeared again and again, sometimes verbatim:
“I wish there was an MPAA rating system for books.”
It comes from a sense of mild absurdity. Movies have ratings. Video games have ratings. Television has ratings. None of those industries collapsed when content information was standardized — they got more accessible, not less. Readers find it bizarre that books, the medium with arguably the most varied internal content, are the only one without any of this.
And it’s not coming from a place of wanting books restricted. The romance community is famously sex-positive, anti-censorship, and protective of authors’ freedom to write what they want. The ask is much simpler:
“I’m as anti-censorship as they come. But there are books with very graphic depictions of horrible things that you really have to be in a good headspace for, and knowing what you’re about to encounter isn’t censorship.”
Telling a reader what’s inside a book doesn’t ban the book. It doesn’t remove it from a shelf. It doesn’t shame anyone for reading it. It respects the reader enough to let them make their own choice.
What We’re Building
We started Swoon Shelves because romance readers deserve a discovery tool that matches the sophistication they already bring to reading.
We don’t rely on users to tag books. We use language models to read and analyze every romance novel systematically — generating consistent metadata for tropes, spice level, content warnings, pace, heat curves, and tone. Every book gets the same depth of analysis, whether it’s a #1 bestseller or a Kindle Unlimited indie release with twelve reviews. That solves the “coverage drops off for less popular books” problem that has hobbled every community-sourced platform, including the ones romance readers love most.
Spice levels are first-class data. Closed-door, kissing only, fade-to-black, moderate heat, explicit, erotica — these are real distinctions that change the reading experience completely. We surface them prominently, not as buried tags. You should know before page one.
Content warnings without spoilers. The hardest part of warnings is doing them honestly without ruining the book. Our approach: flag content categories (assault, infidelity, on-page death, pregnancy plot, etc.) without revealing where in the story they appear or how they resolve. You’ll know what’s in the book. You won’t know how the plot turns. Both readers — the one who wants to avoid and the one who doesn’t want spoiled — get respected.
Trope search that actually works. Include the tropes you want. Exclude the ones you don’t. Filter by spice. Filter by length. Filter by “has third-act breakup” or “no third-act breakup.” Romance readers have been asking for this for a decade. We’re building it.
The Right Book for the Right Reader
We believe every romance deserves to be read by its right audience. Not its broadest audience. Not its most marketable audience. Its right audience — the readers who will love it for what it actually is.
A five-pepper dark romance isn’t “worse” than a closed-door inspirational. A morally gray hero isn’t “worse” than a cinnamon roll. They’re different experiences for different readers in different moods. The job of a discovery tool isn’t to judge — it’s to match.
Where the publishing industry has to optimize for sales and social platforms have to optimize for engagement, we get to optimize for something different: fit.
The reader who asked “is there rape in this book, yes or no” deserved an answer. We’re building the tool that gives it to her.