公開ロードマップ

Non-fiction chapters are built on material found on the web right now — real studies, articles, figures and quotes — instead of only what the model happens to remember. Every claim comes back with the source it came from, so you can check it, keep it or swap it. Write about something that changed last month and the book already knows about it.

Having separate author profiles for different pen names would help when writing across multiple genres. Supporting up to 5 pen names per account would allow writers to keep the associated details distinct as they branch out into new genres.

I would like a Google Docs-style version history for each chapter, so that I can switch between previous revisions and avoid losing my content if I make a mistake or an AI generation fails.

Allow creating a named series and adding books to it once. Currently, creating a new book in a series requires re-importing every previous book or selecting them individually from the full library, then manually re-editing characters and locations, which risks inconsistencies across the series.

Checking consistency while writing currently requires copying each chapter into an external document, because the application does not allow any export before the book is complete. Being able to discuss and analyze the content produced so far would allow revising and correcting issues without waiting for the end of the manuscript.

Allowing authors to specify word count ranges for chapters or sections helps control pacing, estimate effort and costs, and meet editorial standards. Setting target lengths during planning reduces revisions and ensures the content matches length guidelines.

Character portrait cards help with marketing and can be used as a book cover maker when added to a character page. Having a tool to generate them would help show off characters and their stories.

I would like to make manual text edits directly in the Book Editor without having to select the text and invoke the AI each time.

See rank, reviews and sales for every book you have published in one place. Each book comes with one specific suggested fix — a weak blurb, a wrong category, a missing keyword. You spend your time on the book that actually needs it.

Allowing authors to upload real or synthetic voice samples would help maintain personal authenticity across audiobooks. It would also enable multi-character voice recordings for fiction, allow quick audiobook updates after manuscript edits, and let an author's distinct voice narrate books in multiple foreign languages.

Add an option to use a custom OpenAI-compatible API key as a bring-your-own-key (BYOK) provider.

Convert children's books, teen books, adult books, and e-books into flipbook-style videos that can be uploaded and watched on YouTube and social media.

Moving the editor to the step directly before the publishing section within Book Flow would make the entire writing and publishing process much more streamlined.

Add an optional button to automatically generate an image for a chapter so there is no need to write a prompt manually.

Vocallab.ai is a more affordable alternative to ElevenLabs. I would like to be able to connect my own Vocallab API key directly to the app for audio generation.

Locations are an essential part of a story codex alongside characters. When bringing in a previous book or importing an outline that already includes locations, the system ignores them all. Having to re-enter every location manually can take up to an hour for projects with large settings.

It is essential to have strong security features to protect accounts and information. The platform should offer two-factor authentication (2FA) that does not rely on email as an option for all accounts.

A find-and-replace tool would make it easy to edit and swap out names, phrases, and other recurring terms without having to manually read back through each chapter.

It would be helpful to chat directly with the AI about uploaded reference material and past chapters. When writing a long series, it is easy to forget earlier details or notes, so having a dedicated chat tab to query uploaded resources or ask for feedback on a specific chapter would save time.

Add a search magnifying glass or find feature directly on the chapter editing page so writers can quickly locate a specific passage in their text.

Allow visual resources and authoring tools to use the active book's context directly—such as pulling in its title, author, and text elements automatically—instead of requiring us to re-enter details manually.

Include a dark mode because it is easier on the eyes, and many creators and writers prefer high-contrast colors and less visual strain. Additionally, when using a tablet or laptop, dark mode helps reduce battery consumption.

Every image comes out at the size and under the rules the platform actually enforces: BookBub's 300x250 static frame with your cover filling about half of it, Amazon Sponsored Brands headlines inside the 50-character limit with no price or savings claims, Sponsored Products inside 150, and Meta sets in both 1:1 and 4:5. Get any of that wrong and the ad is rejected, or it runs and quietly wastes the spend. You get several variants per book, so you can test which one earns the click instead of running one image and hoping.

Add a visual thumbnail for each cover generation style option. Users currently spend credits experimenting to see what each option produces, because there is no preview showing what each choice looks like.

A built-in recording setup with a scrolling teleprompter would help authors record their own audiobooks directly in the app. Basic editing tools to stop, start, clip out mistakes, and balance audio levels across sessions would make it easy to get started, with the option to expand capability via Web Audio plugins.

Allow chapters to be organized and grouped into acts or distinct narrative sections to help structure the story.

I have notes, visuals, and concepts for my universe, including characters, locations, and objects. It would be helpful to upload images to keep visual references alongside the notes for characters, items, and settings while writing.

Currently, many AI wrapper tools feel disconnected, requiring a lot of repetitive work across separate steps. The platform should instead offer an integrated AI assistant that guides the writer through the entire process, from analyzing market gaps to publishing the finished book.

Add the ability to delete uploaded files inside the editor when they become redundant, and add folder support to better organize uploaded files.

There are times when using BYOK is preferred and times when it is not. Currently, switching between the two requires storing the key elsewhere and repasting it, deleting and recreating the key, or disabling it with the provider. An in-app toggle to turn the key on or off, along with granular control over which operations use the key, would make this workflow much easier.

Adding a button to turn a character description into an image within the novel editor would make it easier to visualize the character and adjust details to match the story's world.

After using the book editor to make changes, I would like an option to import or save the updated version back into My Books directly.

Generating images is credit-intensive and happens automatically, but first-pass images are rarely the final choice, leading to extra credit usage. Giving users the option to choose whether or not to generate images would prevent wasted credits.

Offering a REST API and Model Context Protocol (MCP) support would allow developers to automate their workflows and integrate with external tools.

The editor frequently resets the font and size while editing, requiring the user to constantly fix typography as they work, which significantly slows down the workflow.

While authors strive for uniqueness, readers often look for and recognize familiar cover styles for specific genres. Providing cover templates would make it easier to design covers that immediately capture reader attention.

Allow more control over manga and comic layouts, including typesetting and font formatting. A blank canvas where creators can visually customize their layout, choose the number of panels, and arrange them—similar to how Anifusion works—would give much more flexibility.

I would like the Trend Detector to include filters for market, country, and language.

A scene is the true building block of a story: one point of view, one setting, and one clear shift. When generating an entire chapter at once, the AI decides the scene breaks on its own and tends to blur them together. Writing one scene at a time keeps the generation focused, prevents a single weak section from ruining the whole chapter, and makes it much easier to reorder beats when the pacing needs adjustment.

I would like an option to download the complete editorial review report. The Review section provides valuable analysis covering structure, characters, chapters, style, and specific editorial comments, but currently only the original and edited manuscripts can be downloaded. A dedicated "Download Review Report" option, ideally exporting to PDF or a similar format, would make it easier to consult the feedback while revising, compare versions, and keep the report as part of the editing workflow.

Prologues currently cannot be added until the Format stage, which remains locked until the book is written and the metadata is set up. This creates extra work from having to renumber and fix chapters after the fact. Allowing writers to plan and write the prologue during the outline and chapter stages would make the workflow much smoother.

When a character's name is changed, that update should apply everywhere across the entire book, or the editor should include a book-wide search-and-replace tool.

When continuing a series, the system should extract key information from uploaded earlier books to automatically populate the world and character profiles. Writers could then review and edit the extracted profiles to ensure accuracy, making worldbuilding across a series much smoother.

Integrate feedback from the Review stage directly into the manuscript editor during Style Correction. Rather than leaving recommendations in a separate editorial report, mark the exact paragraphs and dialogue they refer to with in-text highlights, margin annotations, or comments indicating what to shorten, expand, rewrite, delete, or keep. This would let writers use the tool as an editorial assistant and revise systematically without searching manually for each issue raised in the report.

Pick a page up and drop it straight where you want it, instead of stepping it one place at a time with the buttons. Moving a page to the far end of a long storyboard stops being a dozen clicks.

In the picture book feature, the contents page currently lists every single page, often as 'Page 1, Page 2...'. Allowing authors to group pages into named chapters or sections—so only those named sections appear in the table of contents while unnamed pages are skipped—would make the contents page much cleaner.

Currently, enabling the Table of Contents inserts the full page brief into each TOC entry. Having a dedicated, optional TOC-title field per page would avoid pulling from the text-on-picture field, which is used for lettering instructions rather than titles. In fixed-layout EPUB picture books, the navigation document serves as both the reader's jump menu and the accessibility tree, so allowing named landmarks such as 'Title', 'Exhibits', or 'Index' is much more useful than a plain numbered page list.

Many non-fiction ideas do not need to be full-length books; they work better as lead magnets for a book series or other offers. Please consider adding a lead magnet generation and formatting feature, similar to Gamma.

Currently, the visual style of the Gemini model is very recognisable and does not always meet expectations. Letting writers select among different image-generation models would provide better variety and results as models evolve.

I would like to be able to edit the text directly using the information from the editor report, and then export the finished manuscript as a PDF or formatted book file in addition to DOCX.

Currently, character details are set once and reused throughout the book. If a character changes over time, their profile must be manually updated after each chapter, which overwrites earlier state information and disrupts context when editing previous chapters or writing flashbacks. The product should dynamically track entity states (characters, locations, items) chapter by chapter—either automatically inferred by the AI or guided by custom entity rules—while preserving version history so earlier chapters retain their proper context.

Add an optional "Generate Complete Book" button alongside the current step-by-step workflow to generate an entire draft in one action once the description, world, characters, and storyboard are set up. To avoid unexpected credit deductions, show a confirmation screen displaying the total page count and the exact credit cost before proceeding. After generation, users can review the whole draft in one pass and selectively regenerate specific pages, chapters, or images that need adjustment. This reduces friction for rapid testing, first drafts, and visual or children's books.

A trilogy or series typically uses very similar book covers while changing small details such as color, objects, or tone as the series progresses. Let the cover generator use previously created covers as a reference to produce matching styles for subsequent books, with an option to specify exactly what should change.

Code blocks are used extensively in both programming texts and LitRPG books. Having proper code block formatting in the editor would make writing and formatting these books much easier.

Spread mode renders one wide picture that runs across the left and right pages and the gutter between them, instead of two matching pictures, one per page. Landscapes, big crowds and the reveal moments of a picture book fill the whole open spread the way they do in the books on your shelf.

Backing up an in-progress book gives peace of mind halfway through writing. Being able to export the completed chapters as a simple .txt file would provide a straightforward backup.

Collaborating with another writer currently requires copying, pasting, and re-importing chapters manually. It would be helpful to have native co-writing features, such as multi-seat accounts sharing a credit pool or guest access where collaborators can draw from their own credits or the host account's credits.

Reordering pages is currently difficult, particularly when inserting a new page at the very beginning of a project. Having a simple visual drag-and-drop page overview grid would make reorganising pages much easier.

Adding voice-to-text input directly to text boxes would make it easier to capture details while ideas are fresh.

During chapter generation, targeting the AI editor at a specific passage currently requires describing that passage in the prompt. Highlighting a section of text and having a prompt pop up directly for that selection would allow edits to apply specifically to the selected words.

While characters can currently be added and locations can be stored alongside them, locations are important enough to warrant their own section in the story codex. Setting brings stories to life and requires different details and fields than characters do, so having a dedicated locations section makes sense.

Please add support for Swahili translation, as it is an official language in my country and the primary reason for my purchase. Thank you.

It would be extremely helpful to upload a text-only PDF of a book and have the software automatically parse the scenes, generating complementary illustrations alongside or within the existing text layout.

We need the ability to fully design back covers and spines: adding and removing elements, using AI to generate visuals that match the front cover, and styling the spine beyond a simple solid color.

A dedicated course module would provide writers with a clear, structured learning path to develop their skills, rather than having to rely on scattered external resources. Similar to tools like GAJIX, it could offer personalized learning paths and direct guidance.

Allow audiobook narration with third-party TTS providers beyond ElevenLabs using our own API keys. Most newer providers offer OpenAI-compatible endpoints (such as Xiaomi MiMo-V2.5-TTS, which supports voice cloning from a short clip without a transcript, voice design from text descriptions, and inline emotion tags like (sighs)). Adding a provider dropdown with an API key field and a live connection check, similar to the existing ElevenLabs integration, would work well.