DockCat

How To Use This File

This file records the ChatGPT prompts we used to generate DockCat art assets: default cat characters, app icons, and collectables. Feel free to use them for generating the cat identity, postures, animation frames, and collectables to customize your DockCat. If you would like the cat to look more realistic or more cartoon-like, you can adjust the art style part of the prompt in the way you like.

Contents

• Prompts > Cat: shared cat art style, static posture prompts, and animation frame prompts.
• Prompts > App Icon: sleeping and non-sleeping Dock icon source prompts.
• Prompts > Collectables: prompts for collectables the cat can bring home.

How To Use Prompts

For cat assets, start with Cat > General Art Style, then combine it with either the static posture prompt template or the animation frame prompt template. Replace the bracketed placeholders with the posture, motion plan, frame count, and output paths you need.

For one-off assets, you can use the example prompt as a tested starting point. The examples omit file-format requirements already covered by the templates, so add the template requirements when generating production images.

For custom cats, upload reference photos before prompting. The prompt assumes ChatGPT can see those references and should preserve the cat’s identity: coat color, markings, fur length, body shape, facial features, and overall personality.

Prompts

Cat

General Art Style

Use the uploaded real cat photos as reference images. Transform the cat into a cute retro web-game style 2D digital illustration while preserving the cat’s original coat color, fur length, markings, body shape, facial features, and overall identity exactly as shown in the reference photo. Do not change the cat into a different breed, fur type, or coat pattern. The cat must be in the posture or animation frame specified by the production prompt.

Render the cat in a simplified, rounded, cartoon-like style inspired by early 2000s browser pet games or Flash game animal icons. The overall visual style should feel soft, charming, nostalgic, and slightly low-detail, like an old web UI pet illustration or sticker asset. Use clean, smooth outlines with slightly thick dark linework, simple flat or softly blended colors, and minimal gentle shading. Suggest the fur texture and markings using a few simplified curved lines, short strokes, or stylized pattern shapes rather than realistic individual hairs.

Keep the anatomy cute and stylized, with a compact rounded form, slightly enlarged head proportions, small paws, and a gentle, appealing expression. The final result should look like a small illustrated pet icon from an old casual browser game: clean silhouette, warm and slightly muted colors, soft hand-drawn digital painting feel, low visual complexity, and a nostalgic 2000s pet-game aesthetic.

The cat must be isolated with a fully transparent background. There should be no background elements at all: no floor, no shadows on the ground, no scenery, no room, no gradient backdrop, and no decorative objects. Output only the cat as a clean standalone transparent-background character asset / sticker-style illustration.

Static Pose Prompts

Prompt Template

Create a single static transparent-background PNG of the referenced cat in this posture: [Describe the exact cat posture, body orientation, expression, and any app-specific use case here.]

Use the art style specified above. Preserve the cat’s exact identity, coat markings, fur length, colors, face shape, and body proportions from the reference photos.

File requirements:

• File format: PNG with alpha transparency.
• Color space: sRGB.
• Default canvas: 1024 x 1024 px.
• Runtime display target: the app preserves each PNG’s source aspect ratio and renders poses at 15% of the source PNG point size by default. Users can adjust this scale percentage.
• Padding: keep at least 8% transparent padding around the cat.
• Anchor consistency: keep the cat’s body-bottom or paw-bottom anchor visually consistent across all poses in the same asset pack.
• Naming: use lowercase English letters, numbers, underscores, and hyphens only.
• Background: fully transparent. No shadow, floor, scenery, object, UI, text, or decorative element.
• Output file: [Add the target path, for example poses/resting/side.png.]

Rest Side

Create the referenced cat lying on its side in a relaxed resting pose. The cat should look calm and comfortable, with a soft rounded silhouette and visible tail placement.

Example output file: poses/resting/side.png

Rest Loaf

Create the referenced cat in a loaf pose with paws tucked under the body and eyes closed. The cat should look relaxed, compact, and gently alert.

Example output file: poses/resting/loaf.png

Held

Create the referenced cat being held by the armpits, shown as a cute dangling desktop-pet pose. The cat’s body may hang slightly downward with small paws visible and the body naturally stretched longer due to gravity, but the expression should remain gentle and charming rather than distressed. Do not draw human hands or arms.

Example output file: poses/held/held.png

Dialogue Stand

Create the referenced cat standing or sitting upright facing the viewer directly, suitable as one concrete dialogue pose for reminder dialogs and returning-home messages. The expression should be attentive, friendly, and slightly expectant. Leave enough transparent space above the head in the canvas for the app to place a speech bubble. Do not include the speech bubble.

Example output file: poses/dialogue/stand.png

Transition Stretch

Create the referenced cat at the peak of a satisfying stretch, suitable as one concrete 2-second transition pose: front paws stretched far forward, body low in front, back arched, and rear raised upward. This is not an animation.

Example output file: poses/transition/stretch.png

Animation Frame Prompts

Walking Animation Tips

Walking frames are usually harder to make consistent than static poses. If your AI image generation tool cannot generate multiple frames in one request, try generating the walk frames one at a time. You can also try the staged workflow below instead of asking for unrelated single frames:

1. Generate one canonical static side-view image of the cat first. This image is the identity reference for the walk cycle.
2. Generate the walk cycle as one strip or one planned set, using the canonical side-view image and the original cat references in the same request.
3. Keep only one walking direction. DockCat mirrors the frames in the app for the opposite direction.
4. Review the frames as an animation before adding them to the pack. Reject frames with size popping, baseline jumps, identity drift, copied guide marks, cropped paws/tail, shadows, speed lines, dust, floor patches, or inconsistent padding.
5. If only one or two frames are wrong, regenerate those frames with the canonical side-view image plus the neighboring correct frames. Do not restart the whole pack unless the cat identity changed across the full cycle.

When your image tool can generate a single contact strip, prefer this layout because it gives the model one shared scale and baseline:

• 4 equal square cells in one horizontal row.
• Same transparent or plain removable background in every cell.
• The cat centered in each cell with the same visual size.
• Paw-bottom baseline aligned across all cells.
• No visible borders, labels, frame numbers, arrows, guide marks, floor, shadow, or scenery in the final image.

After generating a strip, crop or export each cell to separate PNG files named walk_01.png, walk_02.png, walk_03.png, and walk_04.png. Every exported frame should keep the same canvas size.

Prompt Template

Create a transparent-background PNG animation cycle of the referenced cat performing this motion: [Describe the animation action, movement direction, mood, body mechanics, and intended app use case here.]

The cycle should contain [frame count] frames. Before generating images, plan the motion for every frame so the cycle connects smoothly:

• [frame_01.png]: [Describe the pose for frame 1.]
• [frame_02.png]: [Describe the pose for frame 2.]
• [Add or remove frame rows as needed.]

Use the art style specified above. Preserve the cat’s exact identity, coat markings, fur length, colors, face shape, and body proportions from the reference photos across all frames.

File requirements:

• File format: PNG with alpha transparency.
• Color space: sRGB.
• Default canvas: 1024 x 1024 px for every frame.
• Runtime display target: the app preserves each PNG’s source aspect ratio and renders animation frames at 15% of the source PNG point size by default. Users can adjust this scale percentage.
• Padding: keep at least 8% transparent padding around the cat.
• Frame consistency: use identical canvas size, identical cat scale, consistent baseline or paw-bottom position, consistent head/body size, consistent line weight, consistent color palette, and consistent transparent padding for every frame.
• Naming: use lowercase English letters, numbers, underscores, and hyphens only. Sortable filenames are required because the app plays every loadable image in filename order.
• Background: fully transparent. No shadow, floor, scenery, object, UI, text, or decorative element.
• Output folder: [Add the target folder, for example animations/walk/.]

If generating frames one at a time, always reference the planned full-frame cycle and the previously generated frames. Keep canvas size, scale, anchor position, line weight, color palette, and visual style identical across all frames.

Walk Frames Prompt

Create a looping walk cycle of the referenced cat walking to one side. Before generating images, plan the motion for every frame so the cycle connects smoothly:

• walk_01.png: first contact pose in a natural four-legged walk, with one front paw reaching forward and the opposite rear paw pushing back.
• walk_02.png: passing pose that continues from Frame 1, with the body moving forward over the planted paws and the lifted paws swinging through.
• walk_03.png: opposite contact pose, using the other diagonal pair from Frame 1.
• walk_04.png: second passing pose that continues from Frame 3 and leads naturally back into Frame 1.

The walk should be slow, relaxed, and cute, suitable for a desktop pet strolling along the macOS Dock. Let the model solve the exact paw placement as a natural cat walk, but make sure the front legs and rear legs both alternate across the loop. Keep the body anchor stable so the app can move the character horizontally.

Example output folder: animations/walk/

Walk Strip Prompt

Use this when your image generation tool can produce one image containing all walk frames:

Create one horizontal 4-frame walk-cycle strip of the referenced cat walking to the right, suitable for DockCat’s animations/walk/ frames.

Use the uploaded canonical side-view cat image as the primary identity reference, and also use the original cat photos as supporting references. Preserve the cat’s exact coat markings, fur length, colors, face shape, body proportions, line weight, palette, and cute retro web-game style across all 4 frames.

Layout:

• One image containing 4 equal square cells in a single horizontal row.
• Each cell represents one frame of the same looping walk cycle.
• The cells must share the same cat scale, same paw-bottom baseline, same canvas padding, same lighting, and same transparent or plain removable background.
• Do not include visible cell borders, frame numbers, labels, arrows, guides, floor, shadow, scenery, UI, or text.

Motion plan:

• Frame 1: first contact pose in a natural four-legged walk, with one front paw reaching forward and the opposite rear paw pushing back.
• Frame 2: passing pose that continues from Frame 1, with the body moving forward over the planted paws and the lifted paws swinging through.
• Frame 3: opposite contact pose, using the other diagonal pair from Frame 1.
• Frame 4: second passing pose that continues from Frame 3 and leads naturally back into Frame 1.

The walk should be slow, relaxed, and cute, suitable for a desktop pet strolling along the macOS Dock. Show motion through paws, legs, a tiny body bob, and subtle tail position changes only. Let the image model choose natural paw placement, but check that both the front legs and rear legs alternate visibly, Frame 2 follows from Frame 1, Frame 4 follows from Frame 3, and Frame 4 loops cleanly into Frame 1. Do not draw speed lines, dust clouds, motion trails, ground shadows, floor patches, detached effects, or extra props.

After generation, export the 4 cells as separate PNG files named walk_01.png through walk_04.png, keeping identical canvas size and alpha transparency for every frame.

Walk QA Checklist

Before using generated walk frames in DockCat, check:

• All frames have the same pixel dimensions.
• The cat’s visible size is consistent across frames.
• Paw-bottom or body-bottom baseline stays aligned.
• Coat markings, face, tail, and body proportions do not drift.
• No frame contains floor, shadow, scenery, text, guide marks, or detached effects.
• The cycle reads correctly in filename order and loops cleanly from the last frame back to the first.
• The app only needs one walking direction; do not add separate left and right folders.

App Icon

Sleep State App Icon

Create a macOS app icon source image showing the referenced cat sleeping peacefully on a cute square cushion. The cushion should be soft, simple, rounded-square, and charming, with a retro web-game pet-app feeling. Preserve the cat’s exact identity, coat markings, fur length, colors, face shape, and body proportions from the reference photos. The cat should be curled up or compactly sleeping on top of the cushion. Use the shared cute retro 2D illustration style, but compose it as a polished macOS app icon. The background outside the icon artwork should be transparent or app-icon-safe, with no text, no extra objects, and no scenery.

File requirements:

• Output file: icon_sleep.png.
• Custom pack path: app_icons/icon_sleep.png.
• 1024 x 1024 px PNG source.
• Square app-icon composition.
• Keep enough padding for macOS icon masking and scaling.
• This source should be exportable to .icns.
• For bundled runtime Resources/AppIcon, downsample the transparent PNG to about 512 px on the long edge after generating the high-resolution source, then inset the visible artwork to about 86% of the transparent source image.
• DockCat caches successfully loaded custom icons under ~/Library/Application Support/DockCat/AppIcon/; regenerate the pack source and restart DockCat when changing the icon.

Active State App Icon

Create a macOS app icon source image showing only the same cute square cushion from the sleeping app icon, with no cat on it. The cushion should match the sleeping icon exactly in style, shape, color palette, angle, scale, and lighting, so the Dock icon can switch between the cat-sleeping version and the empty-cushion version without feeling like a different icon. Use a soft, charming retro web-game pet-app illustration style. The background outside the icon artwork should be transparent or app-icon-safe, with no text, no extra objects, and no scenery.

File requirements:

• Output file: icon_empty.png.
• Custom pack path: app_icons/icon_empty.png.
• 1024 x 1024 px PNG source.
• Square app-icon composition.
• Keep enough padding for macOS icon masking and scaling.
• This source should be exportable to .icns.
• For bundled runtime Resources/AppIcon, downsample the transparent PNG to about 512 px on the long edge after generating the high-resolution source, then inset the visible artwork to about 86% of the transparent source image.
• DockCat caches successfully loaded custom icons under ~/Library/Application Support/DockCat/AppIcon/; regenerate the pack source and restart DockCat when changing the icon.

Collectables

Prompt Template

Create a single small transparent-background PNG collectable item for DockCat: [Describe the specific object, silhouette, angle, color, condition, and any story detail here.]

Use a visual style compatible with the DockCat cat assets: simplified, rounded, charming, nostalgic early-2000s browser pet-game style, with clean smooth outlines, slightly thick dark linework, warm muted colors, flat or softly blended fills, minimal gentle shading, and low visual complexity. The collectable should feel like a tiny object the cat could bring home from a walk. It should be readable at small size and visually compatible with the cat.

If you add a display name for a new collectable, keep the English name at 16 characters or fewer, including spaces.

File requirements:

• File format: PNG with alpha transparency.
• Color space: sRGB.
• Default canvas: 128 x 128 px.
• Padding: keep at least 10% transparent padding around the object.
• Naming: use lowercase English letters, numbers, underscores, and hyphens only.
• Background: fully transparent. No shadow, floor, scenery, UI, text, or decorative element.
• Output file: [Add the target path and filename.]

Leaf

Create a small fallen leaf collectable, slightly curled and simple, with a warm muted yellow-green or soft orange color. The silhouette should be readable and cute rather than realistic, with one or two simplified vein lines and a gently rounded shape. It should look like a tiny object the cat proudly brought home.

Example output file: collectables/leaf.png

Pebble

Create a small smooth pebble collectable with a rounded irregular oval shape, soft gray-blue coloring, and one or two subtle lighter markings. The pebble should look tactile, simple, and charming, like a tiny keepsake from outside, while staying low-detail and readable at small size.

Example output file: collectables/pebble.png

Feather

Create a feather collectable, light and gently curved, with a soft cream, pale gray, or muted tan color. The silhouette should be rounded and friendly rather than sharp, with a simple central shaft and only a few broad feather barbs so it remains readable at small size. It should feel like an ordinary outdoor bit the cat proudly carried home.

Example output file: collectables/feather.png

This site is open source. Improve this page.