Threadline Travel Memory Poster

Threadline Travel Memory Poster

Use the system imagegen skill and image-generation tool. Edit the supplied photographs; do not invent unrelated people, poses, places, actions, or scenery.

Output contract

• Produce exactly one independent raster poster per uploaded photo. Never combine photos or outputs into a collage, contact sheet, diptych, or shared canvas.
• Use a 3:4 portrait composition with an upper photographic region occupying about 58–66% and a lower warm-paper region occupying about 34–42%.
• Build a negative-form memory structure, not a normal photo-above/illustration-below comparison: the people are precisely removed into warm paper-white voids above, then the same action returns as smaller solid charcoal echoes below.
• Preserve every source person's count, location, scale, pose, action direction, overlap, contact relationship, and core carried object.
• Connect both versions with one source-meaningful, hand-made continuous line and readable text arranged along its path.

This treatment works best for photographs containing one to three fully visible people with a clear directional action such as walking, holding hands, carrying an umbrella, supporting one another, gliding, embracing, climbing, or pulling. Prefer sources with a road, railing, shadow, waterfront, rope, track, bridge perspective, slope, or motion trace that can support the connecting line. If the source is less suitable, adapt from its strongest truthful gesture or environmental direction without inventing a false event.

Read each source before editing

Privately determine for each photo:

1. every person's exact silhouette, position, scale, pose, movement direction, overlap, touch point, and essential object;
2. the real place, perspective, light, weather, spatial structure, and dominant environmental colors;
3. one meaningful action or environmental cue that can become the continuous line, such as joined hands, umbrella handle, camera strap, footsteps, road curve, bridge perspective, railing, glide path, waterfront, or cast shadow;
4. the best lower placement—left, right, or lower center—based on the source action direction;
5. one source-grounded English title of 1–3 words, one short field label such as FIELD NOTE 01, and one original observation sentence of 5–12 English words;
6. the exact 2–4 text fragments made by splitting that observation sentence without changing or adding words.

Determine the observation sentence verbatim before generation. Do not let the image model improvise alternate wording, fake letters, or extra copy.

Preserve the environment and remove the people above

Keep the upper region recognizably photographic and faithful to the source. Preserve the architecture, road, landscape, water, sky, light, shadows, perspective, weather, location colors, depth, materials, and spatial relationships. Apply only slight desaturation, powdery softness, restrained grain, and the color character of an aged art publication.

Replace each person at the exact original position with a warm paper-white negative silhouette. The void must cover the original photographed body precisely, including its original scale, pose, action direction, overlaps, contact points, and essential objects such as an umbrella, backpack, camera, or carried item when those objects belong to the action silhouette.

Use a clean edge with subtle paper fiber. Add only a few restrained graphite interior marks when needed to identify hair, hands, clothing divisions, backpack, umbrella, camera, or a contact point. The interior remains paper white: no photographic color, skin, clothing texture, background leakage, translucent fill, or white-outline-only treatment.

The result must feel as though the people physically left the real scene and accurate paper-white spaces remain. Never move, enlarge, beautify, simplify into a generic fashion silhouette, or change the pose of a removed person.

Return the same action below

Make the lower region one continuous field of warm ivory, coarse-fiber paper. Recreate the same group as smaller solid echoes in deep charcoal gray, faded black ink, or graphite. Preserve exactly the same number, poses, action direction, contact relationships, overlap logic, and core objects as above.

Keep the lower figures visibly smaller than the upper voids. As a group, they should usually occupy about 20–42% of the lower region's width and 25–55% of its height. Place them left, right, or lower center so their position answers the upper action vertically, diagonally, or along its movement direction.

Leave at least 55% of the lower paper visibly blank and continuous. Add only one to three very faint graphite traces derived from the source—such as a road, slope, railing, waterfront, or directional edge. Do not redraw the full environment, complete a landscape, or fill the paper with decoration.

Draw one meaningful threadline across both layers

Develop the selected source cue into one continuous hand-drawn line in ochre red, dark red, rust red, or gray blue. Prefer a hue that harmonizes with the source and remains restrained.

The line must:

• begin at a real action contact point in the upper figures or their core object;
• follow a truthful environmental or movement direction visible in the source;
• visibly cross the horizontal boundary without stopping;
• connect to, wrap around, or continue from the corresponding hand, footstep, contact point, or core object of the lower figures;
• retain physical pencil, waxed thread, or stitched-line texture with slight wobble, pressure changes, pauses, loops, and intermittent marks.

Do not create a smooth Bézier path, glowing digital stroke, arbitrary flourish, repeated decorative lines, or a line that loses semantic connection to the source action.

Add source-grounded text along the path

Include only:

• one 1–3 word English title;
• one short field label, usually FIELD NOTE 01 or another concise source-appropriate index;
• the predetermined 5–12 word English observation sentence, split into exactly 2–4 readable fragments.

Distribute the sentence fragments across both regions along the road direction, action direction, or main-line arc. At least one fragment must closely follow, touch, wrap around, or visually continue the cross-region line. Preserve the sentence's exact wording and word order across the fragments.

Use thin hesitant handwriting, pencil lowercase, or wax-thread-like lettering. Text may rotate gently along curves but must remain legible. Place the main title in the lower continuous blank paper at a size readable in chat preview. Keep the field label subordinate.

Do not add an author name, signature, logo, URL, QR code, watermark, false place/date, fake quote, extra caption, dense ornament, rigid digital text box, or pseudo-text.

Palette and material

Keep the palette restrained:

• upper region: the source's main location colors with reduced saturation and visual noise;
• upper people: warm paper white with minimal graphite detail;
• lower region: warm ivory paper, charcoal gray, faded black ink, and graphite;
• auxiliary traces: one or two very pale low-saturation colors sampled from the source;
• connecting line: one dark red, rust, ochre-red, or gray-blue accent.

Paper grain, graphite, wax, and stitch-like irregularity should feel physical and matte. Avoid colorful cartoons, polished vector graphics, glossy effects, fashion silhouettes, and dense decorative editorial styling.

Compile one source-specific edit prompt per photo

For every source image, explicitly state:

1. one separate 3:4 output and the chosen upper/lower percentage within the allowed ranges;
2. the exact people count, positions, scales, poses, direction, overlaps, touch points, and core objects that must remain unchanged;
3. the truthful upper environment, restrained publication color treatment, and exact paper-white negative replacements;
4. the lower echo's scale, placement, pose correspondence, charcoal material, minimum 55% blank paper, and one-to-three source-derived environmental traces;
5. the chosen real action/environment cue, line color, physical mark character, precise upper origin, boundary crossing, route, and precise lower connection;
6. the exact title, field label, complete observation sentence, its exact 2–4 fragments, and each fragment's placement and relation to the line;
7. the hard exclusions from this skill, especially no photographic fill inside upper figures, no shifted negative shapes, no changed pose or count, no oversized lower figures, no complete lower background, no stopped or digitally smooth line, no fake text, no logo, and no watermark.

Use concrete source-specific nouns and spatial instructions. Do not rely on style adjectives alone.

Inspect and correct

Review every output at full size and thumbnail size. Regenerate with one focused correction when a core invariant fails:

• If upper figures retain photographic fill, show only white outlines, leak background, or shift away from their source positions, restore exact opaque warm-paper voids aligned to the originals.
• If upper and lower figures differ in count, pose, direction, overlap, contact, or core objects, make the lower echo match the upper source action exactly.
• If lower figures are too large or the lower environment is fully redrawn, shrink and simplify them and restore at least 55% continuous blank paper.
• If the line stops at the boundary, floats decoratively, or looks digitally smooth, reconnect it through both regions from a real upper contact point to the corresponding lower point with physical pressure variation.
• If text is invented, unreadable, out of order, or misspelled, restore the predetermined title, field label, and exact observation sentence fragments.
• If the result reads as a normal split poster, strengthen the defining absence/return relationship: accurate paper-white people-shaped voids above and the same smaller solid action below.

Before returning, verify that every photo produced a separate poster and that one tactile line plus its along-path text binds the exact missing pose above to the same returned pose below as one continuous travel memory.