Community Watch: Royal Road Piracy — Google Play Books Listings & Full-Text Datasets

Community Watch advisory, first published July 15, 2026. Two related concerns affecting Royal Road authors: (1) serialized Royal Road stories appearing as unauthorized ebook listings on Google Play Books, with step-by-step DMCA and evidence-preservation guidance; and (2) a technical risk assessment of a 1.61-million-row chapter-level Royal Road full-text dataset and a reader-comment generation adapter published on Hugging Face — what those capabilities enable, what is verified, and what remains unproven.

Updated July 25, 2026 with Royal Road's official response: new DMCA guidance, a Publication Verification feature, and confirmed takedowns of several fake publisher names.

Community members have reported serialized Royal Road stories appearing as unauthorized ebook listings on Google Play Books. Reports suggest this may be ongoing, so a clean search today should not be treated as permanent confirmation that your work is safe.

What authors should do

1. Search Google Play Books for your story titles, series titles, and pen names.
2. If you find an unauthorized listing, screenshot everything before reporting it: the listing URL, title, cover, description, price, claimed author and publisher, publication date, copied passages, and copied author notes or shout-outs.
3. Save the listing through the Wayback Machine's Save Page Now tool when possible. Listings can disappear and later return under a different URL or identifier, so keep both the live and archived URLs.
4. Use Google's legal troubleshooter to start the copyright-removal request. Select Google Play and follow the prompts for a copyright complaint involving a book; Google can change the exact labels in this flow. Its copyright reporting guidance and book-removal guidance provide additional information.
5. Include the infringing listing URL, every URL where you originally published the work, and a clear description of what was copied. Royal Road fiction and author-profile URLs are especially useful, along with any other dated publication pages.
6. Recheck periodically for new or re-created listings.
7. Keep evidence establishing your publication history, including Royal Road URLs and timestamps, original files, drafts, and other dated records.

If Google asks you to connect your legal name to your pen name

Google may ask for more evidence if the legal name on a complaint is not visibly connected to the pen name on the original work. If that happens, one practical option is to add a temporary, dated verification note to your public Royal Road author profile or another page only you control, then reply to Google's email with that URL.

For example:

Verification note: I, [LEGAL NAME], confirm the copyright complaint submitted to Google LLC on [DATE OF COMPLAINT] regarding an unauthorized edition of [STORY TITLE]. — [DATE POSTED]

Only publish information that is accurate and that you are comfortable making public. Remove the note after Google confirms it is no longer needed, and do not disclose more personal information than the reporting process requires.

Some reported listings have included copied chapter text, author notes, and reader shout-outs. Preserve those pages: material unique to your serial can be useful evidence connecting the unauthorized ebook to your original publication.

Why this matters

An unauthorized listing does not transfer ownership of your copyright, but it can create a serious rights-verification dispute. Amazon KDP's rights guidance says it may request documentation proving publishing rights and warns that rights problems can lead to rejection or removal, account impact, or loss of royalties. Unauthorized ebook availability may also complicate KDP Select exclusivity. Prepare your evidence before publishing or enrolling.

This notice is practical community guidance, not legal advice.

Updated Investigation — July 17, 2026

We reviewed two public Hugging Face repositories that involve Royal Road material. They do different jobs: one packages story chapters and metadata as a bulk dataset; the other is described as a fine-tuned adapter for producing Royal Road-style reader comments. Either can have legitimate research or testing uses. Both also reduce the technical effort required to reuse community writing or simulate community response at scale.

Bottom line

The immediate concern is not that these two projects are proven to be working together. We found no evidence of that. The concern is capability. A large, chapter-level fiction corpus makes copying, searching, sorting, model training, and style analysis much easier than visiting stories one at a time. A comment-generation adapter can produce reactions that may look like reader engagement. Used separately or combined by a third party, those capabilities could support story imitation, unlicensed training, market mining, or manufactured comments.

What we verified in the full-text dataset

The public dataset is OmniAICreator/RoyalRoad-1.61M. Hugging Face's viewer reports approximately 1.61 million rows, and the repository lists 12.5 GB of files in Parquet format. The records are chapter-level, not merely a catalog of fiction titles.

Visible fields include fiction and chapter identifiers, story and chapter titles, author names and profile URLs, cover and chapter URLs, tags, content warnings, descriptions, ratings, views, followers, favorites, page counts, release timestamps, and a text field containing chapter prose. That structure is already suitable for filtering, analysis, bulk export, and machine-learning pipelines. It materially lowers the cost of assembling a Royal Road-specific corpus.

The repository applies an MIT label to the dataset. That label describes the repository's stated license; it does not, by itself, establish that the uploader owns or can relicense every author's underlying story text.

About the dataset uploader

The dataset was uploaded under the public account OmniAICreator. The profile lists interests in LLM and text-to-speech work and shows other large corpus releases, including Japanese-Novels-23M, WebNovels-Ja, Qiita-1.07M, and pixiv-dic-202506. This establishes a pattern of publishing large text collections. It does not establish unlawful intent, author consent, or the exact collection method used for RoyalRoad-1.61M.

What the comment adapter can do

The separate repository is SamuelYo/royalroad-comment-sft. It is described as a supervised fine-tuned LoRA/PEFT adapter, not a complete standalone chatbot. A developer would load it with a compatible base language model.

Given a chapter, excerpt, summary, or prompt, an adapter of this type can generate contextually plausible reader reactions: praise or criticism, character and pacing feedback, predictions, questions, jokes, and emotional responses. Clearly labeled synthetic comments can be useful for interface testing or research. Publishing them as if they came from real readers would create misleading engagement and undermine trust.

Threat assessment

• Full-text dataset — high concern: it centralizes story prose and detailed performance metadata in a downloadable, machine-readable form. Plausible misuse includes bulk copying, targeted imitation, fiction-model training without author permission, and identifying commercially attractive genres, authors, or stories.
• Reader-comment adapter — medium-high concern: it can lower the cost of producing convincing audience reactions. Plausible misuse includes fake social proof, automated comment campaigns, and distorted feedback signals.
• Combined capability — consequential but unverified: a third party could pair story text with synthetic comments, but we found no verified link between these repositories and no evidence that their uploaders are coordinating.

Capability is not proof of misuse. We have not established who collected the chapter text, what permissions were obtained, whether any model has been trained on it, who has downloaded it, or whether generated comments have been posted to Royal Road.

What authors can do now

1. Check the dataset viewer for your fiction title, author name, chapter URL, and distinctive passages.
2. Preserve the repository URL, dataset card, visible schema, matching rows, and timestamps with screenshots and archived copies before filing a report.
3. Compare exact passages against your dated Royal Road pages and retain the original URLs and source files.
4. If your copyrighted text appears and you did not authorize it, review Hugging Face's Content Policy, which provides an intellectual-property takedown route.
5. Report the matter to Royal Road through its official channels and keep a record of every notice and response.

Royal Road's Terms of Service state that authors retain ownership of content they post and restrict scraping and non-personal exploitation. Whether a particular dataset entry or downstream use is legally infringing depends on facts we cannot determine from the repository alone. This update is a technical risk assessment and practical evidence guide, not a legal conclusion.