113End-of-Day Anime Caption

Generate this image now. Draw it directly. Create exactly ONE finished portrait image in this single turn: a single ORIGINAL anime-style scene with ONE subtitle line, ready to share. The user attaches one photo of ONE person in the SAME message. Preserve recognizable facial structure, apparent age, skin, hair and glasses, especially "【Visible feature to keep】" if actually visible. If the detail conflicts, trust the photo. If there is no person, request a portrait rather than inventing one.

Transform the person into an original hand-drawn anime character just home and resting on a sofa in one small cozy room. Their jacket is still loosely on, one bag rests on the floor and one phone lies face-down on a low table. A warm lamp and darkening window suggest evening without claiming an exact time. "【What I got through today】" guides the feeling only. Do not illustrate several episodes or invent a school, workplace, person or illness. This is a single quiet reaction moment, not a comic strip.

Use one fixed analog-cel-animation-inspired look: variable dark outlines, hand-painted flat color, soft shadows, faint film grain, indigo night, amber lamp light and warm beige. Keep coherent room perspective, accurate hands and knees, and a face large enough to recognize at list size. No named anime, artist, studio, character, fake episode label, speed lines or speech balloons.

At 69–78% image height print only "【One subtitle】" as a large white subtitle with thin dark outline over a calm dark patch. If it says "Choose for me", write one short empathetic or lightly funny English line based only on the supplied real context and resting scene. Never print the instruction phrase or context field. Wrap a long line within the subtitle zone without another caption. Leave top 12% and bottom 15% free of essential lettering. No other writing, channel logo, watermark, panel borders, grid, collage, extra image, code or later step. Return only this one finished image.