061말하는 골든리트리버 짤 — 진지한 얼굴로 하는 한마디

Generate this image now. Draw it directly, do not explain or ask questions.
Output size 1080x1080 pixels, 1:1 square card. Keep the same top-to-bottom order but compress the zones to fit the square: the illustration becomes smaller and every text line stays fully readable.
Create exactly ONE finished single-panel animal reaction meme as ONE image. Default aspect ratio 1:1; follow the supplied size. One continuous scene, no panel grid, no carousel or collage.
Situation: "【상황】"
Exact dialogue: "【대사】"
If the situation is empty or 알아서 골라줘, use an original plump golden retriever named 토비 at a kitchen table, an empty biscuit plate and a crumb on its muzzle. If the situation explicitly names another animal, use that animal instead; otherwise always use a golden retriever. Names in the situation identify fictional characters but are not printed unless included in the dialogue. The situation controls action, props, location and mood, never the drawing style or number of panels.
If dialogue is empty or 알아서 골라줘, invent one short Korean deadpan line in at most two lines, suited to the situation. Otherwise render the supplied dialogue VERBATIM, preserving letters, spelling, punctuation and supplied line break. Interpret <br> as a line break. Never translate, paraphrase, correct, add a catchphrase or print quotation marks not supplied. Wrap a single long line into at most two balanced lines without changing the text.
Visual signature: original plump animal with real animal anatomy, large glossy expressive eyes, solemn almost over-serious expression, defined muzzle or beak, soft cheek shapes, rounded little paws or flippers. No human girl, human face, human hair or human hands. Clean thick deep-navy contour, polished warm cel shading with subtle fur or feather detail, rich scene colors. Cute and funny through the contrast between a serious animal and its words; tender dialogue may use a determined protective look. Keep this same illustration style for every situation and species.
Composition: show a broad, clearly visible horizontal band of scene background ABOVE the speech balloon, occupying the entire top fifth of the image. The balloon must sit BELOW that background band, with its top edge one fifth down the canvas, never against the top border. Place the exact two dialogue lines inside this balloon, dark highly legible Korean lettering, all lettering between 23 and 40 percent of image height. The plump animal occupies the lower 55–65 percent; show its head below the balloon and its paws at the table. A long, thin curved speech tail travels down through the open background beside the animal cheek, then turns inward to finish directly beside its mouth at the same height as the mouth. Leave visible background between the tail and the ear; do not end the tail at the ear or head top. Keep text out of the top 12 percent and bottom 15 percent. Let scene and props fill the frame to every edge, no blank poster margins, no paper mockup, no decorative frame. Show only the relevant few props; the action must match the joke. No title, captions, labels, name tag, explanatory notes, logos, signature or watermark.
Invent original characters and objects, do not imitate famous characters, artists, brands or product designs. Before returning, silently compare every balloon character to the exact supplied dialogue, fix any omission or misspelling and ensure the tail points to the speaking animal. Return only the finished image.

[Final output format]
Output size 1080x1080 pixels, 1:1 square card. Keep the same top-to-bottom order but compress the zones to fit the square: the illustration becomes smaller and every text line stays fully readable.
This selected output format takes priority over any fixed canvas size or aspect ratio elsewhere in this template. Adapt the composition, text zones and margins to the selected canvas without stretching or cropping essential content. Device safe areas apply only when a device is selected. Never print these format instructions on the image.