How Melted Crayons TV Works

Type your idea into the Quick Start box (“What story do you want to tell?”) on the homepage, or on your Studio dashboard once you're signed in, or tap its mic and say it. Then:

• Pick your look: an art style, plus Where will this live? (YouTube / Wide, Reels / TikTok, Instagram, Square or Print / Comic), which sets the panel shape.
• Choose the writer: Write it for me drafts the panels from your idea and invents a cast (1 credit; name a saved character in your idea to write them in), or Keep my exact words puts a script or lyrics you paste or upload onto the panels word for word (free, no AI). If your script names its speakers, each one becomes a role to cast.
• Read your script: see every panel before any art exists. Edit the title, looks, scenes, captions and lines, move panels, join or split them, or delete one, all free. Rewrite this panel costs 1 credit. Happy with it? Looks good, continue is one click.
• Cast it: fill each role with your own character, a free Community Classic or a new one generated on the spot. The writer describes each person it invents, and that description fills in the form when you create them. Or skip a role, and it's drawn from its description instead.
• Draw your first panels: you're offered up to your first 3, and the button shows the credits before anything is spent (Draw my first 3 panels · N credits). Drawing keeps going if you close the tab. Or skip it with Not now, open my script in the Studio. If you're out of credits or haven't verified your email, your story is saved and the step tells you what's next. The story opens in the Studio, where one line shows your next step, such as Generate All · up to N credits, to draw the rest. (Exact-words panels arrive with your lines in place but without a scene, so give each one a scene first.) Once every panel is drawn, that line offers Add voices · N credits if your panels have lines, then Watch or download your story, which opens its story page.

Want an empty canvas? In What are you making?, open a format and pick Start blank to keep its shape, style, starting resolution and who can open it, or choose Something else? Start a freeform story, which starts at Standard (1K). A blank story opens in the Studio with four ways to begin: Write with AI (the Story Writer), Start with a blank panel, Browse templates, or Start from my video or photo, which adds a panel and opens its Use a video row, where Use my photo starts from a photo instead. You can change the shape, style and resolution any time in Settings.

Go to Studio → Characters → + New Character (or hit + New in a panel's Cast tab). Choose a clear photo of yourself, your kid, your pet, anyone you want to star in your stories.

• Best results: a front-facing or three-quarter photo with the face clearly visible
• Avoid: sunglasses, heavy shadows, multiple people in one shot, low resolution images
• Photos automatically resize before upload, so high-res phone photos work fine
• The AI will extract the subject onto a transparent background, and you'll see a preview before saving

In the character creator, switch the source toggle to Generate from Description. Type what you want, like “a young girl with curly red hair, freckles, green eyes, wearing a yellow raincoat”, and the AI produces a clean reference portrait you can use as the character's identity anchor. (You can dictate the description too; see Talk instead of type.)

• Best for: fictional or anonymous heroes: dragons, robots, space pirates, cartoon kids, made-up characters of any kind
• Honest tradeoff: real photos preserve identity slightly better across many panels. Generated characters are still very consistent, just not pixel-perfect for real-world resemblance
• Click Try Again to regenerate if the first result isn't quite right. Each generation uses 1 credit.
• The generated image flows through the same background-removal step as uploaded photos. Once extracted, the character behaves identically downstream

Whether you uploaded a photo or generated one, an Adjust this image box sits under the reference. Type the change you want (“give her a red jacket and round glasses”) and the AI edits that one thing while keeping the character on-model. It beats starting over: a fresh Try Again rolls a brand-new result and loses what you liked, while Adjust keeps the parts that already work.

• Keep or Discard: the adjusted version shows as a candidate. Keep it to make it the reference (it re-runs background removal automatically) or Discard to leave the original untouched. A bad edit can never cost you a good image.
• Works on saved assets too: reopen a character, product, brand, or location later and adjust its reference the same way, without recreating it.
• Each adjustment uses 1 credit, the same as a generation.

The moment you upload a photo or generate one from a description, the AI starts extracting the subject onto a transparent background, with no extra button to click. You'll see “Removing background” under the Save button while it runs. The extracted cutout shows next to the source when it's done; the Save button activates as soon as both are ready.

• Runs in your browser when it can: fast, and free. If your browser can't run the model, we finish the cutout on our servers instead (still free), so you can always complete the character. Your photo is stored with your character when you save it.
• First time you do this, the AI model downloads (~30MB, cached forever after). Subsequent extractions are 1–3 seconds.
• If the result looks off, click Retry next to the preview to re-run against the same source.

Give your character a name. For people-style characters you can also set a default pose (standing, action, sitting, running, flying) and default expression (neutral, happy, angry, etc.). They apply whenever the character is cast: the default pose pre-fills the panel's Pose & Action box, and you can type something different there for any one panel. The Usual outfit field says what they normally wear (see Outfits).

Voice selection also lives on the character (see below) so the same character speaks with the same voice across every story they appear in.

Every character automatically gets a distinct speaking voice, matched to their gender when you set it, so they talk in their own voice the moment they're cast, with no setup. Want a specific voice? Pick one and it overrides the auto-choice everywhere that character speaks.

• Filter by voice type: a quick toggle row above the dropdown (All / Feminine / Masculine / Neutral) so you can narrow to the timbre that fits your character
• Grouped by gender, then age: Feminine and Masculine sections each split into Young, Middle-aged, and Older, with Narrators and Specialty (villains, heroes, creatures, announcers) in their own groups
• Each option shows the voice's accent inline (e.g. Bright Australian · Australian); the roster spans American, British, Australian, South African and more, balanced across feminine and masculine so every character finds a fitting voice
• Click the play icon next to a selected voice to preview it before committing
• Auto-assigned, not silent: a character with no hand-picked voice still speaks: we give them a distinct, gender-fitting voice (and keep different characters sounding different). Set the character's Gender on its design page to guide the match.

Don't want to make a character from scratch? We ship a curated set of Community Classics: public-domain legends like Dracula, Alice, the Monkey King, Cinderella and the cast of Oz, drawn in our own original style and free for everyone to use. Look for the gold ★ star on Community Classics in the character picker (hover the star for its label).

• Use one instantly: cast a classic straight into a panel, same as your own characters. It already has a voice and defaults set.
• Make it your own: hit Add to my characters to copy a classic into your library, then edit its name, look, traits, or voice however you like. Your copy is private and fully yours.
• Every classic is an original public-domain reimagining, not a copy of any studio's design.

Characters aren't limited to people. You can also create Products (like a coffee cup or a sneaker) or Brands (a logo or mascot). The AI preserves their exact appearance the same way it preserves a face. Useful for product storytelling, ads, branded narratives, or storyboarding a commercial before you shoot it.

Heroes don't have to be human. Upload a character design for a pet, a dragon, a robot, a creature, or any supernatural being. Then tick Non-human character on the character form. That one flag tells the AI to preserve the subject's exact species, anatomy, and design from your reference instead of nudging it toward human features and proportions.

• When to tick it: any time the character isn't a person: an animal, monster, alien, mascot creature, or fantastical being. Leave it off for human (or human-like) characters.
• Why it matters: without the flag the AI applies its usual human-anatomy guidance, which can quietly straighten a creature toward a humanoid shape. The flag keeps a four-legged dragon four-legged and a robot mechanical.
• Upload or generate: works with both an uploaded character sheet (great for a comic artist's original design) and a generated-from-description creature. Background removal is tuned for the non-human subject too.

Melted Crayons TV ships with 30+ art styles: comic book, anime, watercolor, toy bricks, hand-painted anime, claymation, pixel art, noir, pop art, cyberpunk, and more. Each style has its own visual language, and the AI is trained to render in it consistently. The style choice often defines the format too: ink and halftone reads as a graphic novel, photoreal as a storyboard, comic-book as a Saturday-morning strip.

You don't have to guess what each one looks like: the Art Style picker shows a sample image of every style, rendered on the AI engine you've selected, so you preview the look before spending a single generation. Switching engines swaps the previews, since the same style has a different feel on each.

• Storybook pick: “storybook” for bedtime stories or “comic” for classic sequential art
• 🖍️ Coloring Book: generates every panel as clean black-outline line art on white, and because captions and bubbles are overlays, the PDF download of a Coloring Book story is a printable coloring book of your own characters (PDF is on Storyteller and up). Write it together, print it, hand over the crayons. To bind one, use the Print-ready PDF. Selling it takes a plan with commercial use, which starts on Creator.
• Show-stopping pick: “Toy Bricks” or “Hand-Painted Anime”: these styles are instantly recognizable and feel premium
• Brand / pre-viz pick: “cinematic” or “photorealistic” for product marketing or storyboarding a shoot before you film it
• You can change the style mid-story, but the AI works hardest to stay consistent when you keep one style throughout

Beyond the built-in styles, you can hand the AI a style reference image: a comic panel, an illustration, a character render, anything whose artistic treatment you want your story to echo. Open Settings → Style Reference in the Studio header and upload your reference. The AI reads the image and writes a description of its look (linework, palette, shading, texture, mood) and applies that to every panel.

• It copies the style, not the subject: the AI describes the treatment in words (“bold ink outlines, muted palette, hard shadows”) and paints your scenes that way. It does not copy the people, objects, or composition from your reference.
• It takes precedence: a style reference overrides the automatic look the AI would otherwise learn from your first panel, so it's the strongest way to lock a consistent feel from panel one.
• Best references: a single clean panel or a character render reads more clearly than a busy collage. You can Replace or Remove the reference any time from the same Settings panel; removing it reverts to the auto-extracted style.

Each panel can pick a secondary style to blend with the story's primary art style. Pick “watercolor” as your story style, then on a dream sequence add a blend with “Hand-Painted Anime”, and the panel comes out as a watercolor base with its soft hand-painted feel layered in. The AI does the reconciliation; you just pick the two styles you want.

• Where it lives: on the Scene tab of any panel, open the Art style section and use the “+ blend with another style” picker: the same visual gallery as the main style picker, so you preview the blend before choosing. Pick one → re-generate the panel to see it applied. Choose None to clear it; the panel goes back to the primary style on its next render. A blend changes this panel's image only: restyling a video uses your story's look.
• Per-panel, not per-story: blending is a panel-level choice. Use it on a dream/flashback panel without disturbing the rest of the story's look.
• Apply tip: changing the blend only takes effect the next time you generate the panel. AI generations cost credits, so the toggle never silently triggers a re-render.
• Good blend ideas: “noir + comic” for a hard-boiled detective vibe, “storybook + watercolor” for soft children's scenes, “cinematic + cyberpunk” for film-style sci-fi.

By default every panel reads as its own frame, top to bottom. When you want a real comic-book page (several panels laid out together with gutters), open Layout from the Studio header. It's an optional composition mode: arrange panels onto pages for the reader and the printed PDF, without touching how the story animates or narrates.

• Pick a page template: single, 2- and 3-panel strips, 2×2 and 3×2 grids, manga layouts, or a splash-plus-two. Tap a panel from the tray to drop it into a cell, and drag inside a cell to reframe what shows.
• Auto-layout flows every panel into pages of one template in a click, the fastest way to a full page comic.
• Additive, never destructive: panels stay the atomic unit, so Watch Mode, animation, and Listen keep reading the flat panel sequence. Any panel you don't place still appears after the composed pages, so nothing gets left out.
• WYSIWYG: the composer renders real bubbles, comic words, and framing scaled into each cell, so a page looks the same in Layout, in the reader, and in the exported PDF. The Print-ready PDF keeps your pages too.
• Reading direction: the Left to right / Right to left button in the Layout toolbar sets how pages read, for manga-style stories. It mirrors where the cells sit on the page, never the art or the lettering, and the reader, PDFs and comic reader file all follow it. In print, a right-to-left book is bound on the right.

Pick your image resolution per story in Settings → Image Resolution, right under the AI Engine: Standard (1K), High (2K), or Ultra (4K). On the default Nano Banana engine they cost 1 / 2 / 4 credits an image; the other engines render at their own size for a flat 1 credit. A freeform story starts at Standard: crisp for web, social and reading. Stories made from a format or starter start at High, for full-screen video and large displays, and the print formats (books, comics and cards) start at Ultra so their art prints sharp; you can change any of these in Settings. Ultra is on paid plans, for print and hero shots (on the free plan, print formats draw at High). Whether your video download carries a watermark, and your commercial-use license, are set by your plan. Video export quality also steps up with your plan (1080p → 1440p → 4K). More on that in the Save as Video section below.

Open any story and the Studio is laid out around the panel you're working on. A panel is one illustrated frame of your story:

• Panel list (left): every panel in story order. Add panel, Move up / down and Delete live here, plus M to mute a panel (it's hidden from the reader, Watch, Listen and every download, but kept), the star in the tile's bottom corner to make a panel the story cover, and Listen to hear it. A ▶ badge marks panels that have a clip. Near the top of a tile, a violet star marks an on-model look and Off-model? marks a panel to look at (see Keeping your cast on-model).
• The frame (centre): the selected panel at its real shape. Once it's animated, its clip plays right in the frame with your text on top. Under it are two tool tabs: Image (Generate Panel, Refine, Zoom & Flip, Undo / Redo) and Clip (Animate, Lip-sync, Motion transfer, Restyle, your takes and Download). The image tools and Clip open once the panel has art; until then they stay dimmed, and tapping one says what it needs.
• Whole story (under the tools): the actions that work on every panel at once: Story Writer, Generate All (shown while panels have a scene but no art yet, with the most it can cost: “Up to N credits”; it becomes Stop during a batch, which keeps going if you close the tab), Animate All (works once two or more panels have art) and Check continuity (free, works once a panel that isn't muted has both art and a cast). Until then each one stays on screen, dimmed, and says what it needs.
• Inspector (right): four tabs, Cast, Scene, Dialog and QR. Dialog opens once a panel has art or text; QR once it has art. Tap a locked tab and it tells you what to do first. On smaller screens the panels run in a strip and these tabs sit in a bar along the bottom.
• Header: Languages, Settings (art style, style reference, AI engine, image resolution, comic theme, lettering, default aspect ratio), Layout, Share and Finish & publish.
• Find a tool: the first of the tools on the strip above your panels, after the Write, Generate, Animate and Publish steps, searches every tool by what you want to do, like “dance”, “lip-sync” or “QR”. It names the tool, says what it needs if it isn't ready yet, and takes you there. Nothing runs from the list: a paid tool still shows its price before you start it.
• Shortcuts: [ and ] switch panels; ⌘ / Ctrl + Enter generates the selected panel.

Generate All in the Whole story card draws every panel that has a scene but no art yet, and the button shows the most it can cost before you start. It runs on our servers, so it keeps going with the tab closed, and the button shows how far along it is (“2 of 5 · panel 3”).

• Stop: the button becomes Stop during a batch. The panel being drawn finishes, and no more start.
• One bad panel doesn't end the batch: a panel that fails or is turned down is noted and the rest carry on. The summary at the end says what happened (“Generated 3 of 5: 2 failed…”).
• What does stop it: running out of credits, needing to verify your email or reaching a usage limit halts the batch, and the Studio tells you why.
• One batch at a time: across all your stories, one runs at a time, and each takes up to 300 panels. If one is going on another story, stop it there or let it finish first.

The scene prompt is a short description of what the panel shows. Write it like you'd describe a movie shot to a friend.

“A dragon flies low over a glowing forest at dusk. Mist curls between the trees. A castle rises in the distance.”

• Be specific about lighting and mood: “dusk,” “rain-soaked,” “cozy interior,” “harsh noon sun”
• Describe the camera: “low-angle shot,” “close-up,” “wide establishing view”
• Don't describe the characters here: the AI handles them via your reference photos. Just describe the world they're in.
• Or start from Scene presets (an enchanted forest, a city rooftop at night, a space station and more) instead of writing your own

Recurring settings deserve the same treatment as recurring characters. Build a backdrop once in Studio → Scenes & Locations: upload an image or generate one from a description (“a neon-lit alley at night”), name it, and it joins your library. Then on any panel's Scene tab, open The Scene and under Or start from hit Upload a photo or scene to upload a fresh image or pick one from your saved library, to drop that backdrop in. The AI composites your characters onto it instead of inventing a new background.

• Consistency for free: use the same saved scene across multiple panels (or episodes) and the setting stays identical: the environment counterpart to character identity.
• Upload or generate: bring your own backdrop, or let the AI create one. Generated scenes cost 1 credit per attempt; uploads are free.
• Pick once per panel: a chosen scene skips AI scene generation for that panel and becomes the backdrop your cast is composited onto. Clear it to go back to an AI-written scene.
• Adjust a backdrop in place: an Adjust this image box lets you tweak a scene in words (“make it night”, “add fog”) while keeping the rest of the composition intact, with the same Keep / Discard safety as character references.

AI is one way to fill a panel, not the only way. If you draw, photograph, or film, your work enters the same one path as everything else:

Your image IS the panel. After uploading an image to a panel's Scene tab, hit ✓ This is my finished panel art on the preview card. The panel instantly counts as rendered: it appears in the reader, the cover picker, Watch, and every download, no AI pass involved. One click reverts it to a backdrop if you change your mind. Perfect for illustrators, photographers, and bringing hand-drawn work into an animated story.

Open the Scene tab's Camera row to direct the shot without writing it into your prompt:

• Cinematic look: a complete lighting recipe (Golden Backlight, Grey North, Window Light, One Bulb, Sodium Night, Neon, Void and Paint Pop) with a palette you can tweak (+ Custom palette). A look overrides Lighting.
• Shot: close-up, medium shot, long shot, establishing, etc.
• Camera angle: low angle (heroic), high angle (vulnerable), Dutch / canted (off-balance), eye level (neutral), POV
• Lighting: golden hour, Rembrandt, rim light, silhouette, blue hour
• Mood: joyful, tense, melancholic, triumphant

These add cinematic language to your prompt without you having to write it manually. Especially useful when you want a panel to feel like a movie still, not a generic illustration. On panels the AI Story Writer drafts, Shot and Camera angle arrive filled in, and Lighting and Mood do when the moment calls for them. Keep them or change them.

To keep a panel consistent with another, open Reference a frame, point it at another frame and choose Its look (same characters and art style, fresh scene) or Its setting (same place and light, fresh action). Re-generate the panel to apply any of these.

Open the Cast tab and tap a character card to drop it into the panel (or + New to create one). You can:

• Add up to 4 characters, products or brands per panel
• Describe what each character is doing in Pose & Action (“swinging a golf club, back turned”) and their face in Expression (“grinning”, “furious”)
• If a character has a default pose set, it pre-fills Pose & Action (type over it for this panel)
• Pick an Outfit and a Model sheet view for this panel, when the character has named outfits or a model sheet
• Drag to place: until a panel has art, its cast shows on the frame, and dragging someone to the left, center or right tells the AI where to put them. A speaking character's spot counts even if you never moved them, so their bubble's tail points the right way. It's a hint rather than a rule: what your scene text says wins.

The AI uses each character's reference photo for identity preservation, so they'll look like the same person across every panel even though the AI is generating fresh images. With two or more characters on a panel, each reference is labeled with that character's name, so the AI knows who is who.

Roles without a character: anyone the AI writer invents and you don't cast keeps a written look: age, face and skin, build, hair, wardrobe and colors. They're drawn from it in every panel they appear in, and listed at the top of the Cast tab, where you can change the look, or the action and expression for one panel. A role can also become a real character whenever you like.

• The more the look says, the closer they stay from panel to panel. For a face that holds best, cast a character with a reference image.
• Panels already drawn keep their art until you redraw them, so changing a look or casting a role shows on the next render.

The Dialog tab is speaker-driven: pick who's talking at the top and the editor below shows only their lines. Choose the Narrator to write captions, a character to write their dialogue (added lines are auto-assigned to them), or Comic Words for words like POW and BOOM. One speaker at a time keeps it tidy, with no hunting through a mixed list to find who said what.

Two kinds of text overlay:

• Captions: narration boxes, usually placed at the top or bottom of a panel. Style options include narration (default), thought, action, and location stamps.
• Dialogue: speech bubbles tied to a character. Choose between speech (default), thought (cloud-style), shout (loud emphasis), and whisper (italic).

For both, you can dial:

• Position: drag anywhere on the panel. Until you do, a dialogue line sits in the upper part of the panel, above the figures and over its speaker, clear of your other lettering.
• Width: 10–90% of panel width, which controls wrapping
• Font size, color, and family: pick from a set of comic-lettering fonts (clean comic body, handwritten, marker, bold shout caps, typewriter) chosen to read well in bubbles and caption boxes. They look identical in the editor, on the published page, and in exports. Pick Story font to follow the story's Lettering setting.
• Corner radius: sharp rectangle, soft round, or full pill, to match the comic-book style you're going for. Use story style puts a bubble back on the story's Lettering setting.
• Tail direction (dialogue only): tails aim at the speaker on their own. To change one, pick from eight positions (the four edges plus the corners), and Aim at speaker points it back. Thought bubbles use a trail of shrinking circles instead of a pointed tail.

Comic Theme (Modern or Classic): in Story Settings (the ⚙ menu), pick the lettering look for the whole story. Modern (the default) is the sleek graphic-novel style: dark translucent caption boxes, clean bubbles. Classic is the traditional comic-book look: cream narration boxes, comic-yellow action stamps, white balloons with thick black ink and a hard shadow, and a bold yellow shout burst. It restyles every caption and bubble at once, and it carries through identically to the reader, the PDF, social images, and the video export, so what you see is what you download.

Lettering: just below Comic Theme, set one look for the whole book: a font, Size (Small, Standard, Large), Speech bubbles (Classic, Round, Square) and Captions (Square, Rounded). Size applies to every line. The font and shapes fill in only what a line leaves unset, so a bubble you styled by hand keeps its look. It carries through to the Studio, the reader, PDFs, images and the MP4.

Hit Listen on a panel (on its tile in the panel list, or under Narration in the Dialog tab) to hear its captions and dialogue read aloud in each character's voice. The first listen costs 1 credit; replays are free until the words or voices change.

Add classic comic words from the Comic Words section. Pick a style, type your word, and it drops onto the panel. Then drag to place it, and tune the size, angle, and colors. They are drawn on the art and make no sound.

• Six styles: Impact, Electric, Whoosh, Soft, Kapow, and Heavy, each with its own lettering, burst shape, and palette.
• Your word, your colors: type any text and recolor the fill and burst per word (or keep the preset palette).
• Place it precisely: drag to position, then slide for size and rotation; angled words look great breaking across the action.
• Free: comic words don't use a generation credit, and they burn into every export (PDF, social, and video) exactly as you place them.

Animate All sits in the Whole story card under the frame. It works once two or more panels have art, and says so until then (the Animate step on the journey strip opens it too). It opens the shot list: every panel with its motion beat, ready to render as one batch on one engine.

• Direct all clips: pick a motion style (Subtle, Cinematic, or Dynamic) and the AI director looks at each rendered panel (the actual image, plus its scene and dialogue) and writes a beat per clip with real film grammar: every beat names its camera move and its speed, matched to the panel's tone: slow push-ins on quiet emotional moments, quick pans and tracking on action. Free, no credits. Edit any line before you shoot.
• Honest pricing up front: the shot list shows exactly what the batch costs before anything is submitted, and if you're short, offers to animate just the first N.
• One engine, one look: the whole batch renders on the quality you pick, three clips at a time, so consecutive panels share a consistent motion character. If that engine's safety filter turns down a panel, the batch moves to the quality's backup engine for the rest (same price, still one look from there on) and the shot list says so. Panels with an up-to-date clip are left alone unless you tick Include panels that already have a clip (re-animate). Clips made from earlier art are selected by default (the shot list says how many it will replace), so untick any you want to keep.
• Your performances are protected: lip-sync clips are never replaced by a batch (redo those one at a time), and your own footage or restyled clips stay put unless you tick Replace my footage / restyled clips with AI motion.
• Walk away, then retry what failed: renders finish on the provider's side, and you're only charged per clip that completes. If a clip fails, its row says why, and Retry N failed runs just those panels again.

First generation rarely lands perfectly. In the Image tab under the frame you have a few ways to iterate:

• Regenerate: a full fresh render with the same prompt. Good if a panel just landed weird and you want a fresh attempt.
• Refine: open the prompt box and type a text edit instruction (e.g. “make the dragon larger,” “change the lighting to morning,” or “give her a different pose”), then hit Apply and the AI edits the panel without redoing it from scratch. Refine handles any change (scene, lighting, composition, or character pose) in plain words.
• Zoom & Flip: crop in, reposition, or mirror the image without re-rendering it. Free, and the framing carries into the reader, Watch and every download.
• Undo / Redo: every regeneration and refine is saved. Step back through previous versions or step forward again with the arrows.
• If a render fails, the tools under the frame say so and give the reason, such as “A character on this panel has no image to draw from. Add one to the character, or take them off this panel.” If the panel already had art, that art stays.

A few small affordances in the Studio that help you see what you've filled in and what each control does:

• The journey strip: the bar across the top of the Studio (Write → Generate → Animate → Publish) always shows where your story is and highlights the next step; every segment is clickable (one that can't open yet says what it needs) and the last one is your destination: your downloads, and publishing with your trackable link + QR. Story tools sit beside it: Voices & audio, Languages and Layout.
• Tab progress dots: small green dots appear next to Cast, Scene, Dialog, and QR when those tabs have content. Quick way to see what's done at a glance.
• Prices and tooltips on the buttons: Generate Panel and Regenerate show their cost on the button itself (e.g. “· 2 credits”), Lip-sync's tooltip states its price, and hovering any action explains what it does (for example Generate: “Build a fresh panel from your scene + characters”).
• Disabled-Generate hint: when the Generate button is greyed out, a small message tells you to type a scene description, pick a preset or upload a backdrop first.
• Story Settings stay tucked away: settings (art style, style reference, AI engine, resolution, aspect) live behind the Settings button so they don't crowd the workspace once you've set them.

In the Scene tab, open Text in the artwork and switch on Render text inside the image, then type the words. This is for text that should appear inside the artwork: signs, posters, billboards, comic words baked into the scene.

• Use this for: a street sign saying “BEWARE,” a billboard reading “GOTHAM TIMES,” a t-shirt graphic, a chalkboard, a book cover in the scene
• Don't confuse with captions, dialogue, or comic words: those are all overlays drawn on top of the panel (and stay crisp/editable). This switch is for words rendered by the AI inside the artwork itself. For a punchy “BOOM” you can move and recolor, use a comic word; use this switch only when the text should be part of the painted scene (a sign, a poster).
• Modern image engines render text fairly cleanly and work best with short phrases (1–4 words)

When you generate panel 2 and beyond, the AI looks at panel 1 to understand your story's exact visual language (palette, lighting, brushwork, mood) and matches it. This is why your story feels like a coherent piece, not a collection of random AI images.

The technical name is style anchoring. It works automatically. Your job is to make sure your first panel looks the way you want the rest of the story to look.

• Regenerate the source panel and the anchor refreshes automatically: if you redo panel 1 to a different look, later panels you generate will pick up the updated style anchor instead of staying locked to the old one.
• Existing panels stay in their original style until you regenerate them. Style changes don't retroactively re-render finished panels.
• Changing the story's default art style (in story settings) clears the cached anchor and warns you; your existing panels keep their look, and panels you generate afterwards follow the updated style.

Your character reference photos are passed to the AI on every single panel where that character appears. The AI is instructed to reproduce their face, features, hair, and skin tone precisely. This is what lets the same hero appear across many panels and styles without slowly morphing into someone else.

The better your reference photo (clear, well-lit, front-facing), the better the consistency. A blurry side profile gives the AI less to work with. A panel is never quietly drawn with part of its cast missing: if the engines can't draw them, the panel fails and says so.

Melted Crayons TV runs on multiple frontier engines. For images, pick your engine in the story settings:

• Nano Banana (Google Gemini): our default. Excellent character consistency and fine-detail rendering, and fast: the best all-round workhorse.
• GPT Image 2: OpenAI's image model. The most creative and detailed, with richer, more original scenes. A little slower (~1 min a panel), so it's the “make this one beautiful” option for hero panels.
• FLUX Kontext: Black Forest Labs' reference-locked engine. Holds a character's identity from one scene to the next and through successive edits, with a distinct, film-like look. A little slower (~30s a panel), so reach for it when consistency across a sequence matters most.
• Seedream 5.0: ByteDance's engine. Broad stylistic range; kept as an alternative and a fallback.

For animation, the engine follows the motion quality you choose per panel in the Animate action. Each level runs on a different frontier model:

• Standard / Pro: Seedance 2.0 (BytePlus): fast, smooth motion for everyday panels.
• Cinematic / Flagship: Kling (via fal): especially fluid, film-like movement.
• Ultra: Seedance 2.5, the finest single-panel animator, for hero shots.
• Omni: powered by Google Gemini Omni, the premium top engine.

It's a per-panel choice, not a story setting, and every paid plan can reach any level. Spend the frontier engines on your best panels and keep the rest lightweight; if a panel feels off, re-animate it or step up a level.

Narration is text-to-speech with natural prosody. When you hit Listen on a panel, the caption text and dialogue text are voiced in order, each line using its character's assigned voice (or your story's default narrator for captions and any dialogue without a per-character voice).

• Cartesia voices: fast, natural, available on every plan including free.
• ElevenLabs voices: premium voice acting with richer emotional range, unlocked on Creator plans and up. Marked in the picker with a lock badge for free / Storyteller users so you can see what's available.
• Filter by gender: a quick toggle row above the voice dropdown narrows to female / male / neutral so you can match a character without scrolling every voice.
• Conversational pauses: the player adds natural beats between clips: a longer breath when a different character takes their turn, a shorter pause when the same speaker continues. Feels like reading aloud, not reading a script.
• Voice leveling: every clip runs through a gentle compressor before playback so a quiet narrator and a shouty villain sit at the same loudness. No more reaching for the volume knob between speakers.
• Lead-in beat: playback waits about eight-tenths of a second before the first word so the start doesn't feel rushed.

Published stories have two ways to enjoy them, beyond reading the static comic: Listen and Watch. Hit Listen to read the comic top to bottom with full narration. The player walks through panels in order, voicing every caption and line of dialogue with the same per-character voices from the Studio. Listen works on every narrated story.

• Page-hold timer: pick how long the player lingers on each panel after the audio finishes: 1, 2, 3, or 4 seconds. Half the time sits before the audio starts, half after, so the listener has a breath to take in the art on both ends.
• Page-flip pause: a longer beat between panels than within a panel, so panel transitions feel deliberate.
• Same voice leveling as the Studio preview, for a balanced listen across an entire story.

If a story has animated panels, a Watch button appears alongside Listen. Watch Mode plays the story full-screen as a narrated animated episode: each animated panel plays its motion clip while the narration carries the audio, auto-advancing panel to panel. Panels you haven't animated show their still (held for the narration), so a partly-animated story still plays start to finish.

• It plays like a real movie. On Storyteller and up, the owner can hit Render movie and we render the whole story into one smooth MP4 that Watch plays, with no seams between panels. It's built exactly like the Animated Video download, with the same options: captions and bubbles are in unless you tick “Clean cut (no text)” in the Download menu before rendering (the voices still carry the dialogue). If you edit the story afterwards, Watch plays the live version until you re-render. Rendering and re-rendering use no credits, including when the page says video lettering has improved since your movie was made. Rendering a movie needs a verified email, and an account can render up to 20 movies an hour. On the free plan, Watch plays live in the browser.
• Bubbles pop on line by line. Watch draws your captions and bubbles, and each one appears as its line starts, in the numbered read order, so nobody reads ahead. A line with no voice appears at a reading pace. When Watch is playing live, the Text on / Text off button shows or hides them; comic words stay. A rendered movie has no such button: its text is set when you render it, with Clean cut.
• Read, Listen, or Watch. The static comic is always there to read; Listen narrates it; Watch brings the animated panels to life. They're three ways into the same story.
• Translations carry over when Watch plays live: switch language and the live player letters the story in that language too, and speaks it once that language's narration is rendered. A rendered movie always plays in the story's original language, whichever language is selected. For a movie in another language, download the Animated Video in that language.
• Share it anywhere. A published story lives at its own link, with Read, Listen, and Watch all built in. Send the link out and it plays wherever you drop it. Melted Crayons TV is where you make the story, not a feed people have to browse to find it.

The audio reads your captions and dialogue in the order they're positioned on the panel: top to bottom, then left to right within a row. There's no separate sequence to manage; just drop each bubble where you want a reader's eye to land first.

• Numbered badges on every non-empty bubble show the read order at a glance: violet for captions, magenta for dialogue.
• Drag to reorder: moving a bubble higher, lower, left, or right updates its number instantly.
• Two bubbles within ~10% of the panel height read as a row; a slight vertical drift between them won't flip the order, matching how the eye actually groups them.
• Empty bubbles (no text yet) don't take a number and aren't spoken.
• The same numbers set the order bubbles pop on in Watch Mode and the video.

Audio is generated on demand and cached, so replays are free: audio streams straight from Cloudinary. A panel is voiced again (1 credit) when:

• You edit the text of any caption or dialogue
• You change a character's voice
• You change the story's default narrator voice

Dragging a bubble only re-orders the lines you already have, at no charge.

Every account starts on the free Sketch plan: 30 credits a month for images, Story Writer drafts and narration, plus a watermarked 1080p Story Video download, and its first animated clip is on us (Standard quality, 5 seconds, on a verified account). Before you verify your email, a free account can spend 15 credits in its first 7 days; verify it to keep creating, to publish and to upload your own clips. Animation, lip-sync, footage restyle, motion transfer, the Animated Video, a watermark-free Story Video, PDF, Print-ready PDF, webtoon strips, comic reader files and Ultra (4K) images start on Storyteller at $9.99/month, and higher plans add more credits, premium voices and sharper video exports. Most paid plans can also be billed weekly, and a weekly plan gets its full credit allowance every week. Compare every plan, and what Restyle, Lip-sync and Motion transfer cost, on the Pricing page.

Generating a single panel involves resizing your reference photos, sending them to the AI, prompt construction, image generation (~10–30 seconds), and post-processing. Most panels complete within 30 seconds. If one takes longer, the Studio says it's still rendering and keeps going. Your panel appears when it's ready, and you won't be charged twice. (Animations have a Cancel button, and a canceled animation costs nothing.)

If a generation fails, you don't pay for it: only successful panels count toward your plan's credits. Now and then the AI model itself turns a panel, script or edit down; the Studio tells you so in plain words, and nothing is charged.

A slow clip keeps rendering in the background and is charged only when it arrives. If an engine never sends it back, the render stops after 25 minutes (45 for a two-speaker lip-sync), with no charge.

Your words and themes are your own. Slurs in a character's mouth, violence, self-harm, crime and dark history are storytelling, and the check leaves them alone. It blocks sexual content only: anything sexual involving a minor, plus explicit sexual acts and named sexual crimes for anyone. Adult-themed romance, intimacy and sensuality are fine.

The check reads panel, motion, lip-sync and refine prompts before anything is charged. The AI models also have filters of their own, which we don't control, so a model can still turn something down.

Your plan gives you one credit balance that refills every month (or every week on a weekly plan) to spend however you like, on images or motion, in any mix. One credit = one Standard (1K) image. Everyday image actions:

• Generating a panel: on Nano Banana (the default engine), 1 credit at Standard (1K), 2 at High (2K), 4 at Ultra (4K, paid plans); a flat 1 credit on GPT Image 2, Seedream and FLUX Kontext. Freeform stories start at Standard; stories from a format or starter start at High, so plan on 2 credits a panel (4 at Ultra for books, comics and cards on paid plans). The Generate button always shows the exact cost.
• Refine (edit a panel with a text instruction): 1 credit
• A Story Writer draft: 1 credit (Keep my exact words is free)
• Generate from Description, Adjust this image, or a generated scene: 1 credit each try
• A model sheet view: 1 credit each to make
• Check continuity: free, and it never redraws a panel
• Panel narration (TTS): 1 credit per panel that gets voiced; replays and unchanged text are free
• Uploading a character photo, a model sheet view, a backdrop, finished art or your own footage: free

Motion (animating a panel or lip-syncing a character) is heavier, so it costs more credits from the same balance, priced by the quality level and clip length you pick:

• Standard (Seedance 2.0): 6 / 12 credits for a 5s / 10s clip
• Pro (Seedance 2.0): 10 / 20 credits for a 5s / 10s clip
• Cinematic (Kling 3.0): 13 / 26 credits for a 5s / 10s clip
• Flagship (Kling 3.0 Pro): 20 / 40 credits for a 5s / 10s clip
• Ultra (Seedance 2.5): 32 / 64 credits for a 5s / 10s clip
• Omni (Google Gemini Omni): 42 / 63 / 84 credits for a 4s / 6s / 8s clip
• Lip-sync: 18 credits per clip (dialogue up to 5 seconds); two speakers 72 credits per clip (up to 9.6 seconds)
• Restyle your own footage: 13 credits per second (3–10 seconds billed)
• Motion transfer: 7 credits per second of footage (3–15 seconds billed)

Also from the same balance: translating is free, and rendering a language's narration is about 1 credit per narrated panel. Publishing is free; it voices any panels that aren't voiced yet at up to 1 credit each, from the owner's balance.

Because everything draws from the one balance, you decide the trade-off: more images, more animation, or any mix. When your plan's credits run out, prepaid credit packs cover the overflow automatically; when those run out too, the upgrade prompt opens.

When your story's ready, hit Finish & publish in the Studio. It opens the story page's Finish & publish panel, where everything lives in one place: preview the full production with Watch or Listen, set your cover, and dial in the soundtrack and narration/music mix, then either download the file or publish a shareable link.

• Downloading is the deliverable. The finished video is yours to post on your own channels: YouTube, Instagram, your site, anywhere. You don't have to publish to download.
• Publishing is optional; it gives you a hosted link + QR. First choose who can open it: Public (anyone; search engines can find it) or Link only (only people with the link; hidden from search). Staff Training Guide, Social Story & Lesson, Greeting Card & Invitation, Personalized Gift Book and Scene Pitch & Storyboard stories start as Link only, and you can switch any time. Publishing needs a verified email and art on every panel that isn't muted. Hit Publish link and the draft becomes a live page. Publishing is free; any panel that isn't voiced yet gets narration at up to 1 credit each from your balance (if credits run short, it still publishes and those panels stay silent until you top up).
• Share link: once published, it gives you the URL to copy, a QR code, the Public / Link only switch, and Move back to draft. That QR opens the page but doesn't count scans until you press Count scans of this QR, which also lets you download it and change its look.
• If we take a story down, it goes back to a draft and can't be published again or deleted unless we review it. Email support@meltedcrayons.tv if you think that's wrong.
• Your mix plays live. Music and narration volume affect the Watch/Listen preview right there (including a music-only cut), so what you hear is exactly what you download.

All files come from one place: the story page's Download menu. From the Studio, hit Finish & publish on a draft (or View live once it's published) to get there; you don't need to publish first. The menu holds:

• Language (once the story has translations): at the top of the menu, it sets the language of every download below it
• Document → Save as PDF (Storyteller and up): a full storybook with a cover page, every panel and its words, using your Layout pages when you've composed them (in the 🖍️ Coloring Book style, this is your printable coloring book). Also here, both Storyteller and up: Print-ready PDF and Comic reader file (.cbz)
• Video → Save as Story Video and Save as Animated Video, with an optional Subtitle file (details below)
• Images → Panel Images (full-resolution PNGs), Page Images when you've composed pages in Layout, Webtoon strip, 800 px and Webtoon strip, 1080 px (Storyteller and up), and Share Panels via… on devices with a share sheet
• Social images: Instagram, IG Story / Reels, TikTok, X / Twitter and Facebook, as images sized for each platform with captions and dialogue burned in
• Audio → Narration (MP3): the whole story's narration as a single audio file, in whichever language is selected

Anyone signed in who opens a published story can use Download: Panel Images, social images, Narration and the Story Video (watermarked on the free plan), and any video can come with its subtitle file; PDF, Print-ready PDF, webtoon strips, the comic reader file and Animated Video need Storyteller or higher. Visitors who aren't signed in can read, listen and watch, and are asked to create a free account to download.

On the story page (draft or published), open Download → Video and choose Save as Story Video or Save as Animated Video. Your browser builds a real MP4 file on the spot: H.264 video, AAC audio, plays in every share target including iOS Safari, no server-side render farm needed. Save as Story Video is your story as a narrated slideshow, every panel a still. If any of your panels are animated, Save as Animated Video (Storyteller and up) is the same animated episode Watch Mode plays, with motion and lip-sync baked into the file. Both ship with full narration; pick the one that fits the platform you're posting to.

• Matches your story's shape: the output frame uses your panels' aspect ratio automatically. 16:9 stays widescreen, 9:16 exports vertical for TikTok / Reels / Shorts, 1:1 for Instagram feed, 4:3 for classic comic-book.
• Free gets the Story Video too: the static narrated video exports on the free plan at 1080p with a small “Melted Crayons TV” corner wordmark. Animated Video and a clean (no-watermark) export are paid.
• Resolution scales with your plan: free and Storyteller export at 1080p (Full HD), Creator steps up to 1440p (2K), and Studio and Agency render full 4K. Higher tiers get a higher bitrate too, so text and bubble edges stay crisp.
• Text on or off, your choice: by default captions and dialogue are burned in using the same styling as the public reader, each popping on as its line starts. Tick “Clean cut (no text)” to export a text-free cinematic version instead (the narration voices still play), great for a film-like share.
• Voice leveling included: the same compressor the audiobook player uses bakes into the MP4, so loudness is balanced across speakers in the final file.
• Encodes faster than real time: a 90-second story exports in a few seconds rather than 90 seconds. The whole pipeline runs offline in the browser, so timing stays locked: audio and panel transitions never drift.
• Browser support: Chrome or Edge on desktop, or an up-to-date Safari. On a browser that can't encode video, the Video section says so; PDF, images and audio still work there.
• Animated export ships your animations: pick Save as Animated Video and the MP4 bakes in motion clips and lip-sync exactly like Watch Mode, no second tool needed. Animated render takes longer than the Story Video (each animated panel is seeked frame-by-frame), but it stays browser-only and offline. Story Video stays the right choice when you want a quick, lightweight, every-panel-a-still narrated MP4.
• Pick the language: the Language picker at the top of the Download menu applies to both Story and Animated. The lettering switches right away, and the voices switch once that language's narration is rendered (Languages in the Studio). So you can ship one MP4 per region (filename gets a -es, -fr, etc. suffix so they stay distinguishable in your Downloads folder).
• Subtitle file: in the Video section, choose SRT (.srt) for YouTube, TikTok, editors or WebVTT (.vtt) for web players, and the file saves with either video, in the language you picked. Each line starts when it's spoken or appears, on the same frame as its bubble. It pairs well with Clean cut. If your browser blocks the second download, the Subtitles for this video card saves it again.
• Add a music soundtrack: drop in a track from your device and it's mixed under the narration: automatically ducked so the voices stay clear, looped to fill the video, and faded out at the end. It plays the same way in the Watch and Listen previews, so you can balance the volume over the story by ear before exporting. The file never leaves your device; it's not uploaded or stored.
• Title & end cards: tick Title & end cards in the Video section to bookend the video with a title card (your story title + creator) and a branded sign-off carrying your call-to-action and a link (meltedcrayons.tv on the free plan; your own link on paid plans). Add a scannable QR to the end card too: once the story is published, Make this QR trackable creates a counted link right there, or paste one from Links & Campaigns. A code on the end card stays on screen for 6 seconds and doesn't fade out, so people can scan it. A link made with the button carries its logo and style when it has your logo. A link you paste or type stays plain. The preview under these fields shows the exact last frame. A stored movie keeps the look it was rendered with until you render it again, and a logo on one can't be deleted until then. To clear it on any plan, use Remove stored movie at the top of the story, and Watch goes back to the live version. Choosing a code's look happens on the Share QR or the Links page.
• Cinematic transitions are baked in: panels cross-fade into each other, still panels get a slow Ken Burns push, and motion clips stretch to fill their moment instead of looping. It's the same short-film feel as Watch Mode, so the download looks like a little movie, not a slideshow.

Melted Crayons runs a remote MCP (Model Context Protocol) connector. Add it to Claude, ChatGPT or Claude Code and sign in with your Melted Crayons account (Cursor and other tools use a personal key), and the agent can add characters and locations from a reference image, write and build panels, generate and refine art, animate panels, lip-sync a character to a short real vocal, drive a character with the motion of a reference clip, voice panels, translate the text, publish, set the story button, make tracked links and QR codes with a channel and client, and edit them later, pull your link + QR stats and export the numbers and build a client report, acting as you. The developers guide names every tool.

• Self-serve on every plan. Your agent acts as you: your own plan, credits and limits apply. See and disconnect connected apps on your account page any time.
• You stay in charge of cost. Your agent can check your balance and quote any paid step before running it, and your plan's credits are the hard ceiling either way. One difference: a refine your agent runs is priced by the story's resolution (1 / 2 / 4 credits), not the Studio's flat 1.
• Done in the Studio: formats and starter templates, creating a character or location from a description, making model sheets and named outfits, on-model stars and Check continuity, Animate All's director and shot list, Layout pages, changing an existing panel's scene or cast, reordering or deleting panels, lip-syncing a panel to its own dialogue, using or restyling your own footage, rendering a language's narration and collaborators.
• Downloads stay with you: the downloadable video is built in your browser. Your agent takes a story all the way to a published page with art, clips and audio; open that page (View live in the Studio) and use its Download menu in desktop Chrome or Edge, or an up-to-date Safari. The Animated Video needs Storyteller or higher.

Melted Crayons TV is a Progressive Web App, so you can install it on iOS or Android straight from the browser. No App Store, no download, no account-on-file: an icon appears on your home screen and tapping it opens the full-screen app experience, no browser bar in sight. It updates itself automatically every time we ship.

• iPhone / iPad (Safari): open meltedcrayons.tv, tap the Share button in the toolbar, scroll down and tap “Add to Home Screen.” The in-app install banner walks you through it the first time you visit on iOS.
• Android (Chrome / Edge): open meltedcrayons.tv and the install banner appears at the bottom of the screen. Tap Install → done. The OS handles the rest.
• Desktop too: Chrome and Edge on macOS, Windows, and Linux all support installing the app via the address bar's install icon, useful if you want Melted Crayons TV in your dock or taskbar like any other app.
• Offline: a brief loss of connection shows a branded offline page rather than a browser error, and stories you've already viewed can stay cached locally for re-reading on the go.

The site opens in light mode. Use the sun and moon button at the top of any page to switch (it's in the menu on a phone), and this browser remembers your choice. Your artwork never changes: what you see on the canvas is what you download, in either mode. Watch and the other players stay dark so your work stands out.

Invite people to co-edit a story with you, like a shared doc for comics. In the Studio, open Share → Invite Collaborators and add someone by email, or copy a shareable link.

• Full editors: collaborators can edit panels, text, and art, and generate: everything except delete, publish, and managing the roster, which stay with the owner. Anything a collaborator generates uses their own credits; narration voiced when the owner publishes comes from the owner's balance.
• Two ways in: an emailed invite (existing users get instant access; an email without an account gets it the moment they sign up), or a link anyone signed-in can open to join.
• Shared stories appear under My Stories badged Shared. A collaborator can leave anytime; the owner can remove anyone.

Here's what we're building next:

• Custom voices: premium narrator and character voices already ship on Creator and up; next is designing your own voice from a recording.