subject_definitions: <Subject 1> is the character shown in <Picture 1>. Its appearance defines the character to insert, including identity, face, head, hair or fur, skin, body shape, proportions, clothing, accessories, and distinctive visual features. Details of <Subject 1>: [Add character description.] smiling dark haired male anime character in orange blue clothing Details of <Subject 2>: [Add character description.] pink haired woman sitting in vehicle wearing a hoody <Picture 1> is the character identity and appearance reference. <Video 1> contains 90 identical frames of a single source image and is supplied through Add Guide. Treat it as a fixed original image, not as a scene to reinterpret or regenerate. placement_instructions: Position: [Specify the exact insertion location.] <Subject 1> behind car door outside of vehicle. <Subject 2> sitting in vehicle. Pose and orientation: [Specify a fixed pose, facing direction, head angle, and gaze.] <Subject 1> looking into car, hands on window glass. <Subject 2> sitting in vehicle. Scale and depth: [Specify size and distance relative to existing scene elements.] Contact and overlap: [Specify supporting surfaces and which existing elements must remain in front of the inserted character.] summary: Add exactly one instance of <Subject 1> from <Picture 1> to <Video 1>. Everything already present in <Video 1> must remain exactly the same. The inserted character is the only addition. Do not replace, remove, move, resize, redraw, restyle, or otherwise alter any existing subject, object, or part of the environment. Repeat the completed static edit consistently across all 90 frames. preservation_priority: Treat every original frame as a locked background. Preserve the original pixels wherever the inserted character does not directly cover them, except for narrowly localized contact shadows or reflections required to integrate the character. These exceptions authorize only the smallest necessary additions. They do not authorize changes to the underlying scene, global lighting, colors, composition, or existing subjects. Adapt the inserted character to the original image. Do not adapt the original image to the inserted character. retention_analysis: <Picture 1>: character_appearance_transfer — preserve the character's recognizable identity, facial features, hair or fur, skin, body shape, proportions, clothing, accessories, and distinctive features. <Video 1>: original_content_locked — preserve all existing subjects, objects, surfaces, textures, lighting, shadows, reflections, colors, perspective, framing, and composition. Retain every existing element in its original position, size, shape, and appearance. detailed_description: Use <Video 1> as the unchanged foundation of the output. Insert <Subject 1> at the specified location without regenerating or reinterpreting the surrounding image. Preserve every existing person's identity, face, expression, gaze, hair, skin, body proportions, clothing, accessories, posture, and position exactly as shown. Preserve every existing object's shape, size, position, orientation, texture, color, and detail. Preserve all text, signs, logos, patterns, and small background features. Do not introduce replacements, corrections, enhancements, cleanup, beautification, sharpening, relighting, color grading, reframing, cropping, or changes in image style. Use <Picture 1> only to determine the added character's appearance. Do not transfer its background, unrelated objects, framing, lighting, or camera angle. Adapt <Subject 1> to the specified pose and the original scene's scale, perspective, lighting, exposure, color temperature, sharpness, grain, and depth of field. Maintain believable anatomy, balance, weight distribution, and contact with supporting surfaces. Add only the local contact shadows, cast shadows, or reflections necessary for the inserted character. Keep all pre-existing shadows and reflections unchanged. Respect existing depth relationships. Where an original subject or object belongs in front of the inserted character, preserve that foreground element exactly. Do not erase, redraw, or deform it to accommodate the insertion. Do not obscure existing faces, subjects, text, or important objects unless explicitly permitted by the placement instructions. If the requested placement does not fit, adjust only the inserted character's placement or scale within the stated constraints. Blend the inserted character's boundaries naturally without modifying surrounding scene details. Avoid cutout edges, halos, mismatched sharpness, floating poses, and incorrect shadows. Keep the completed edit identical across all 90 frames. Maintain the same character position, pose, expression, gaze, appearance, and lighting from the first frame to the last. Do not introduce blinking, breathing motion, lip movement, body movement, hair or clothing movement, environmental motion, camera movement, zoom, transitions, or progressive changes. Prevent identity drift, duplicated characters, merged bodies, malformed anatomy, extra limbs, texture crawling, flickering, and frame-to-frame variation. The intended result is the exact original image with one additional character naturally integrated into it. Everything else must remain unchanged. overall_soundscape: N/A non_diegetic_music: N/A