Changelog

A new carousel format: Polaroid. Each slide keeps a square picture on top and adds a clean white strip underneath, made for your words — a caption, a line of dialogue, a punchline. Pick it when you start a new Instagram carousel, or switch an existing square carousel to Polaroid in the story settings. Your pictures stay exactly as they are and nothing is redrawn; bubbles and text move along with their slide. You can switch back to Square at any time — if anything sits on the white strip, we tell you how much before it goes. The text tool is available on Polaroid slides so you can write straight onto the strip.

A round of editor polish. Three more speech-bubble shapes to choose from, including a speech bubble with its tail on the side and a caption with a long pointer. When a scene has nothing to correct, Generate now goes straight to drawing instead of asking you to confirm a second time. The save status is now a small cloud at the end of the toolbar instead of a label hanging under it, and the toolbars and tabs have a light outline so they no longer blend into the page. Carousel slides are shown as separate sheets with a gap between them, so it is always clear where one slide ends and the next begins. Lists you scroll — bubble shapes, the / menu — always show their scrollbar, and the / menu no longer slips off the edge of the screen.

Images you add to your pages now save reliably. On long stories with many added pictures, new changes could stop saving without any warning, and those images were missing from downloads. They are now stored properly, existing stories have been repaired, and if a save ever does fail you will see a message instead of silence.

Create a character by describing them in your own words. The character form asks for ten separate details, and not everyone thinks about a character that way. A new button in the Characters toolbar opens a simpler window: give the character a name and a role, then just tell us about them — how they look, what they wear, the one detail that never changes. AI reads it and fills in the full character form for you, so all you do is check it and hit Create. If the description is too thin to draw someone consistently — only a personality, say, with nothing about their looks — it tells you what is missing and gives you an example of what to add, and keeps your text so you can simply add to it.

Every icon in the editor's bottom toolbars now tells you what it does. Hover over any button on the Story, Cover, Characters or Props toolbar and a label appears right above it — including the ones that never had a hint before, like the pages overview, speech bubbles, zoom and delete. And the Create Character window no longer closes when you click somewhere outside it by accident, so a half-filled character is not lost to a stray click; close it with Cancel when you are done.

Download your whole story as a single PDF. Export used to stop at one page at a time, which left anyone with a forty-page book downloading forty files and finding something to staple them together with. Open a story, hit Export, and the popup now has two halves: This page is what you already know, and Whole story gives you Download book (PDF) — your cover and every page rendered into one file, in order. It is free and costs no credits. Pages with no art on them are skipped and we tell you how many went in, a portrait cover on a landscape book stays a portrait sheet, and because a long book takes a few minutes the editor counts the pages off as it renders them. Print resolution means a long story can be a large file. The book follows the same watermark rule as every other export. Quick Story is a single page by design, so there is no book to build.

Every art style now draws on the newer, faster image engine, and every style now works from its written definition alone. Two changes that go together. Until now a story could be handed to one of two renderers depending on the style you picked, and the two draw with a slightly different hand — one engine draws everything from here, which means faster panels and a page that holds its line, its colour and its faces from the first panel to the last. And where a style used to show the model a few example illustrations of itself, it now describes itself in words instead. Those examples were finished drawings with their own people, objects and framing in them, and pieces of that were quietly turning up in panels where they did not belong — most visibly in Retro Comic, which kept printing a papery border into the artwork, a second frame sitting inside the one your page already has. Pages you have already made are untouched; anything you generate from now on picks this up.

Fixed an error message that could wrongly tell you that you were out of credits when you were not. If something went wrong while a scene was being checked or generated, the editor sometimes read the failure as a billing problem and showed the top-up screen — even with a healthy balance, which it then quoted back at you. The editor now only says you are out of credits when your balance is actually the reason, and reports anything else as what it is. No credits were ever charged for the attempts that failed this way.

Seven art styles now draw on a newer, faster image engine: Modern Cinematic, American Superhero, Arcane / Cinematic Animation, Live-Action Cinematic, 3D Animated Movie, Webtoon / Manwha and Noir / Sin City. The bigger change is that they use it for the whole page. Wide cinematic strips, tall panels and the other less common panel shapes were being handed to a second renderer partway down a page, which is why a page could come back looking subtly unlike itself from one panel to the next — the same face drawn with a slightly different hand. One engine now draws every panel, so the line, the colour and the faces hold together across the page. Arcane / Cinematic Animation goes further and works from its written style definition alone, with no example illustrations attached at all, which keeps stray people and objects from wandering in out of them. Pages you have already made are untouched; anything you generate from now on picks this up.

Text now sits properly centred inside speech balloons. Every balloon shape was measured, and your lettering is centred on the balloon itself rather than on an invisible rectangle drawn inside it, so the space above the first line and below the last comes out even. It showed up worst on balloons flipped so the tail points the other way, where there was nearly twice as much room under the last line as over the first, and it also tightens the caption boxes across the top of a page. Balloons you have already lettered keep the position they were saved with and pick up the new centring the moment you nudge or resize them.

Two fixes to text on the page. Letters with tails — p, g, y — were having their bottoms clipped when the text sat tightly in a bubble or caption box, which also meant it came out clipped in your exported book. That is fixed, and the fix reaches pages you have already made, so old stories correct themselves with nothing for you to redo. Speech bubbles and captions now also look the same everywhere — in the editor, on your public story page, and in the export. And the page strip along the bottom now draws the text on each page, not just the artwork, so title pages, credits and any page built out of words are no longer blank rectangles when you go looking for them.

A new art style: Live-Action Cinematic. Instead of drawing your story, it renders it — photorealistic faces with real skin, materials that catch light the way real ones do, volumetric light and shallow focus, closer to a film frame or a game cinematic than to an illustration. It is the first style where your characters look like people rather than drawings of people, which makes it the one to reach for when you want a story that feels shot rather than composed — science fiction, thrillers, grounded drama — or when you want to see yourself in the lead. Pick it when you start a new story; it appears first in the style list.

Panels that hold together. A consistency pass that runs across every story and every art style: your characters hold their look from page to page, the objects you define — a pet, a car, a keepsake — now keep their own identity instead of being redrawn from their description each time, and panels with nobody in frame (establishing shots, close-ups, empty rooms) stay in the same visual world as the rest of your book. Nothing to switch on; it applies to every panel you generate from now on. If a panel used to come out wrong no matter how many times you tried, that is the one worth regenerating.

Speech bubbles, redone — adding dialogue is now built around a curated set of speech-bubble shapes: classic rectangles, soft rounded bubbles, thought clouds, and clean caption boxes. Drop one on any panel and your text auto-fits inside it — sized to fit the shape and set in a proper comic-lettering font, so it reads clearly every time and stays crisp on any device. Move it, resize it, flip the tail, and recolour the bubble and its outline. Bubbles work across every format — full graphic novels, Quick Story, and Instagram Carousels — and on story pages your text now lives right inside the bubble.

Quick Story — a new, one-page format built for making something fast and sharing it. When you start a new story, pick Quick Story: you get a single page (no cover, no props to set up), up to three characters, and a one-tap Share that hands off a ready-to-post image plus a link to your story. Choose portrait or landscape and your panel layout right in the editor — change them freely until you generate, then they lock so your page stays consistent. The fastest path from an idea to something you can post.

Change a story's art style from the editor — changed your mind on the look? As long as you haven't generated any art yet (no panels, characters, or props), open Story Settings, find the Art Style block, hit Change and pick a different style. Everything re-derives automatically. Once you've generated anything in the current style, the style locks to keep your whole story looking consistent — clear your panels, characters, and props to unlock it again.

New art style: 3D Animated Movie — the glossy, heart-warming look of modern CGI features like Pixar and DreamWorks. Fully rendered 3D characters, soft cinematic lighting, and richly detailed worlds, with your characters staying consistent across every panel. Perfect for family adventures, comedy, and heart-warming stories. See the post: https://yarnsaga.com/blog/3d-animated-movie-comic-style

Sharper styles, more consistent characters — we upgraded how art is generated across every style: cleaner linework, richer detail, and characters that hold their face and build more tightly from panel to panel. Nothing to switch on — you'll see it on your next story. More: https://yarnsaga.com/blog/sharper-styles-cheaper-noir

Noir / Sin City is now cheaper — this style runs on a new image engine, so each panel costs fewer credits, with the same razor-sharp, high-contrast black-and-white look.

Quick Start — the fastest way from an idea to a finished comic. Open Quick Start from the top bar and there's already a story idea waiting; tweak it, type your own, or roll another. Pick an art style and hit Start — about a minute later you have a complete three-panel strip with consistent characters, opened right in the editor. No blank canvas, no character setup first. Uses about 1 credit. See the new post: https://yarnsaga.com/blog/quick-start-one-prompt-comic

Instagram Carousel format — create comic strips built for Instagram. When you start a new story, pick Instagram Carousel and choose your slide shape: square 1:1 or portrait 4:5. Every slide is sized natively for Instagram, so nothing gets cropped. Each carousel holds 1 to 20 slides — exactly like a real Instagram post — and you can add, remove, and reorder slides at any time. Your characters stay consistent across every slide, just like in a full graphic novel. See the new page: https://yarnsaga.com/comic-strip-maker

Post-ready ZIP export — export a carousel as a ZIP of numbered JPG slides in reading order, ready to upload straight to Instagram from your phone or a scheduler. By default each slide carries a small 'made with YarnSaga' watermark; remove it for 1 credit per post.

Import characters from another story — if you've already created a character in a story with the same art style, you can now bring them into a new story for free. No credits, no re-generation, no AI variance — the exact same character sheet is reused, so the character looks pixel-identical wherever they appear. Open the Characters tab, click the new copy icon in the toolbar, pick the source story, check the characters you want, hit Import. Perfect for series, episodic projects, or shared casts. See the new guide: https://yarnsaga.com/guide/characters/import

Import props from another story — same idea, applied to props. Reuse spaceships, weapons, signature vehicles, or any prop you've generated in another same-style story. Free, instant, identical. Find it as the copy icon in the Props tab toolbar. Guide: https://yarnsaga.com/guide/props/import

Duplicate protection on import — re-opening the import modal shows previously imported items as already imported and locks them, so you can't accidentally import the same character or prop twice into the same story.

Landscape format — when you create a new story you now pick page orientation: classic Portrait for traditional graphic novel feel, or widescreen Landscape for cinematic, screen-first reading. Pages, covers, panel templates and exports all respect the orientation you choose. Pick once at story creation and every page stays in that format.

Background exports — when you download a page as PNG, JPG, or PDF, you no longer wait while it generates. The export runs on our servers in the background while you keep editing. When it's ready, you'll get a notification and the file appears in a new downloads menu at the top-left of the app (next to your credits). Files stay there for 24 hours, so you can grab them whenever.

Downloads menu with badges — a small download icon in the top-left header shows what's happening with your exports at a glance. A pulsing blue dot means something is generating; a green number badge means that many files are ready and waiting for you.

Sharper, faster exports — speech bubbles now flip and resize correctly in downloaded files (a long-standing bug where bubbles always rendered in their original size and orientation has been fixed). Export speed is roughly 3x faster than before thanks to a new server-side rendering pipeline.

Downloaded PNGs no longer get stuck on a stale cached version when you change a panel and re-export — every download now gives you a fresh, up-to-date file.

Automatic Quality Check — every panel image YarnSaga generates is now reviewed by AI the instant it lands. If the result is off — wrong character, broken anatomy, missing details — the system quietly re-generates for free before you ever see the broken version. Zero clicks, zero wasted credits, sharper output by default.

Smarter multi-character scenes — when two or more characters share a panel, YarnSaga now preserves their relative heights, builds, and skin tones across every shot. A tall character stays tall, a slim character stays slim, and the cast no longer morphs into look-alikes between panels.

Characters with quoted nicknames or unusual names (like Tadashi Maeda "Mizuki") are now properly attached to panels — the AI no longer swaps them with other cast members in multi-character scenes

Just type a character's name in a scene description and YarnSaga attaches them automatically — even without using /. Works with full name, nickname in quotes, or first name alone. Use / when two characters share a name to remove any ambiguity

Each mention of a character in a scene now gets anchored to their reference image individually — multiple appearances on the same panel description stay consistent

Undo and redo — Ctrl+Z / Ctrl+Y (or Ctrl+Shift+Z) now works for all canvas overlays: speech bubbles, text boxes, FX stamps, and uploaded images. Move, resize, rotate, flip, delete, add — all undoable, up to 50 steps. Works on both story pages and the cover

Scene sketch upload now passes the sketch directly to the AI as a composition reference — figures positions, camera angle and spatial layout are matched more accurately

AI scene description from sketch no longer includes image dimensions or aspect ratio — panel proportions are always set by the panel settings

31 new fonts — Inter, Poppins, Montserrat, Bungee, Pacifico, Lobster, Press Start 2P, Modak, Knewave and more (55 total)

Text in PNG, JPG and PDF exports now matches the editor exactly — no more font drift between what you see and what you export

Speech bubble text now exports pixel-perfect — line breaks you typed are preserved exactly in PNG, JPG and PDF exports

Image overlays now appear correctly in PNG, JPG and PDF exports

What's New — changelog is now visible in the About panel so you always know what changed

Resize handles redesigned — the entire border edge is now a drag zone (like Figma). Hover anywhere on a side to see the resize cursor and drag freely

Side resize on speech bubbles, text boxes, ribbons and FX — drag any edge to resize width or height independently without locking proportions

Corner handles now always lock proportions — drag a corner to scale uniformly without holding Shift

Image overlay layer — upload a PNG, SVG, JPG or WebP and place it anywhere on the page or cover. Drag, resize, rotate, flip, adjust opacity and control layer order

Image overlays always maintain aspect ratio when resizing from corners

Clicking the canvas background now properly deactivates a text box — it is no longer editable and any text selection is cleared

Speech bubbles, ribbons, and FX overlays can now be resized freely — drag any corner to change width and height independently. Hold Shift to lock proportions.

Added "View Error Message" button on failed generations — click to see exactly what went wrong

Added "View Details" link on error prop and character cards so you can open them without clicking the image area

Error messages now show the actual reason for failure instead of a generic status code

You now get an instant notification when an image generation fails — no more waiting and wondering

Content policy errors now show a clear "Content not allowed" message explaining the issue

Fixed a bug where text, speech bubbles and ribbons would stay in the wrong place after reordering pages

Cover thumbnail is now refreshed every time you publish — no more stale images after changes

Save indicator is now hidden when a story has no pages yet

Support page now shows a note confirming real team replies — not bots

Upload a photo of a real person to auto-fill a character description in one step — no extra buttons

Dialogue is now automatically detected in scene descriptions and saved separately as speech bubble text

Character sheets now always include fully drawn hands (all fingers) and footwear in every view

Generated images are no longer lost if a connection hiccup happens during saving

Font controls for text boxes — choose font family, size and weight for every text element on the page

Speech bubbles, text boxes and ribbons now save automatically and come back exactly where you left them

Each page keeps its own overlays — switching pages shows only that page's bubbles and text

Cover page has its own separate set of overlays

Rotation handles on canvas elements — drag the handle to rotate speech bubbles, text boxes and ribbons freely

Props panel moved to the left side for quicker access while editing

Canvas handles and controls now stay the same size no matter how much you zoom in or out

Art style now determines which AI model is used to better match the expected visual output

Upload a sketch or mockup image for a panel — AI reads the composition and turns it into a scene description automatically

Props now appear in generated scene images — referenced props are used as visual anchors alongside characters

Scene descriptions are now checked for quality and safety before an image is generated

Props system — create reusable objects (vehicles, weapons, ships, buildings) that appear consistently across panels

Upload a reference image for a prop — AI extracts a detailed description automatically

Props can be edited: update the name or description and a new reference sheet is generated

Free text layer — place text boxes anywhere on the page, resize and position freely, double-click to edit

Ribbon overlays — add decorative ribbon banners to any page

Zoom with Ctrl+scroll (or Cmd+scroll on Mac) — plain scroll now pans the canvas instead

Speech bubbles — stamp SVG speech bubbles onto any page, drag to position, resize from corners, flip horizontally or vertically

Pages can now be reordered by dragging them in the page list

Cover image — generate and publish a cover for your story

AI narrative mode — write a page narrative and let AI split it into panel scene descriptions automatically

Photo transform — upload a photo of a real person to create a character based on their appearance

Character sheets are generated in three views (front, side, back) for more consistent results across panels

Story publishing — publish your story to get a shareable public link

Credits system — buy and use AI credits to generate images

Generated images are delivered in real time via live updates — no need to refresh the page