Turn Any Travel Photo Into a Postcard!

The short answer. Upload one travel photo to ChatGPT, paste the prompt below, and you get a 4:3 poster. Your photograph stays as shot across the left 58 per cent of the frame. The right 42 per cent is warm aged paper carrying a small hand-carved rubber stamp of the same place and four lines of typewriter field notes. One photo per run. Print it at 4x6.

Most travel photos never leave the camera roll. They are good enough to keep and not quite good enough to print, so they sit there. The usual fix is to run one through an AI style filter, which quietly replaces the photograph with an approximation of it, and the thing you actually wanted, the record of having been there, goes missing.

This prompt goes the other way. It leaves the photograph alone and gives it a margin: a strip of aged paper beside it, carrying a small stamp carved from the same view and a few lines of notes. The photo does the recording. The stamp does the remembering.

What the prompt actually makes

The frame is 4:3 landscape, split into two regions with no line drawn between them. The left region, roughly 58 per cent, is your photograph. The model is told to keep the subject, the terrain, the buildings, the people, the light and the original colour atmosphere exactly as they are, and to add nothing beyond a restrained grade and a very fine film grain. It may crop for the layout. It may not redraw.

The right region, roughly 42 per cent, is warm off-white paper with visible fibre and matte texture, and most of it is left blank on purpose. Low in that space sits the stamp, small, taking up only about a third of the right side's height. Below or beside it, four short lines of text: the place, a record number, three keywords and the year.

The stamp is the part that does the work, and it is built by subtraction. The prompt asks the model to look at your photo, find the two or three shapes that make the place recognisable, and throw the rest away: the crowds, the parked cars, the rows of windows, the scattered shrubs, the background that could be anywhere. What survives gets carved into two to four ink colours and pressed onto the paper, badly, on purpose.

How to use it

1. Open ChatGPT and upload your photo. Landmarks, streetscapes and landscapes work best, because they have a shape worth carving.
2. Paste the prompt below into the same message, unchanged.
3. Run one photo at a time. The prompt asks for a single poster per image, and uploading four at once is the quickest way to get a collage you did not want.
4. Read the text before you save the file. Image models still misspell place names, and the whole effect rests on that text looking typed rather than generated.
5. Print at 4x6 for real postcard size, or larger if you are framing it.

The rubber stamp travel poster prompt

Copy everything below and send it with your photo. Nothing in it needs editing on the first run.

Create a Rubber Stamp Travel Field Notes Poster from the photo I upload. Output one poster per photo. No collage, no multi-image combinations. COMPOSITION 4:3 landscape. Divide the frame into a left region and a right region, with no visible dividing line between them. LEFT SIDE (about 58% of the frame) Keep the original photo faithfully. Preserve subject identity, terrain, architecture, plants, people, spatial relationships, natural lighting and shadow, authentic textures and the original colour atmosphere. Apply only restrained, art-publication-level colour grading and extremely subtle, fine-grained film noise. Natural cropping is allowed to fit the layout. Do not stretch, distort, shift, replace or redraw the main subject. RIGHT SIDE (about 42% of the frame) Warm off-white aged paper: subtle fibres, natural grain, light usage marks, matte texture. Preserve large areas of unprinted paper. The blank space is an essential part of the layout. THE STAMP Analyse the photo and extract the most location-distinctive elements: subject outlines, architectural structures, terrain contours, plant forms, roads, shorelines or other key visual relationships. Compress them into one small multi-colour rubber stamp image. Do not replicate every element item by item. Retain only the minimum information needed to instantly recognise the original location, subject and scene relationships. Remove crowds, vehicles, dense windows, repetitive buildings, fragmented vegetation, decorative elements and irrelevant background. Place the stamp in the lower-middle of the right-side paper area, occupying only about 30% to 38% of the right region's height, with ample whitespace around it. Do not enlarge it into a standard illustration, a full landscape painting or a brand logo. STAMP ORGANISATION, BY PHOTO TYPE Iconic architecture: retain the most distinctive outer contours, roofs, domes, arches, towers or main structures. Mountain settlements: compress buildings into a few terraced colour blocks aligned along the terrain. Coastal scenery: retain mountain contours, settlement layers, shorelines and sparse intermittent water ripples. City panoramas: retain the main skyline, one iconic building, and one or two layers of distant mountains. Natural landscapes: retain primary mountain forms, trees, shorelines or road orientations. Foreground occlusions: if narratively important in the original photo, retain them as foreground stamp outlines. INK Extract 2 to 4 spot inks from the original photo. Prioritise desaturated colours such as carbon black, deep green, brick red, ochre yellow, slate blue or taupe brown, but do not force a fixed palette. Preserve the most distinctive colour character of the original photo, allowing only a small area of colour for visual emphasis. PRINTING Render each colour as a separately hand-stamped pass: authentic rubber stamp carving texture, hand-engraved marks, uneven line widths, contour notches, fractured edges, dry ink shortages, paper show-through, granular ink, uneven pressure, partial ghosting, and about 1 to 2 mm of subtle misregistration. Allow natural misalignment between colour layers. Edges must not be digitally smoothed. The print should resemble a real carved stamp pressed onto aged paper, not a filtered photo, a smooth vector illustration or a line-art logo. TEXT Generate from the photo's location, theme and visual imagery: Location English name No. number Three short English keywords Gregorian calendar year Place the text below or adjacent to the stamp in the whitespace, in a small, restrained, slightly mechanically imperfect typewriter font. The typography should evoke a traveller's field record, not an ad headline. Spell all text accurately. Add no irrelevant slogans, brands or decorative copy. MOOD Field notes kept by an architect, travel writer or natural observer: quiet, restrained, tactilely real, regionally specific, with handmade imperfections and a collectible feel. The photo handles the on-site record. The stamp captures the most recognisable fragments of memory. AVOID Obvious central dividing lines, circular seals, Chinese red stamps, postage stamp perforations, wax seals, sticker collages, tourist souvenir templates, smooth vector logos, generic city icons, full replication of all architecture, dense detailing, childlike craftiness, cartoon style, 3D rendering, plastic textures, glossy digital gradients, oversaturation, excessive text, decorative clutter, and any redrawing or altering of the left-side original photo.

What each instruction is doing

The prompt is long because almost every line in it is closing off a way the model usually gets this wrong. If you plan to adapt it, it helps to know which line is holding up which part.

The 58 and 42 split, with no line between

Uneven proportions are what make it read as a printed page rather than a template. Ask for a half and half split and you get something that looks like a social media carousel. The instruction to leave out a dividing line matters too, because models love to draw a border, and a border turns two related things into two unrelated ones.

Do not redraw the photo

Left to itself, an image model treats your upload as a suggestion and repaints it: a face shifts, a building gains a floor, the sky changes weather. The list of things to preserve, and the flat refusal to stretch, shift, replace or redraw, is the single most important part of the prompt. If you cut anything, do not cut that.

A stamp, not an illustration

The size cap, 30 to 38 per cent of the right side's height, is doing quiet work. A stamp that grows past that stops being a mark and becomes a picture, and then you have two pictures of the same place sitting next to each other. Keeping it small also forces the model to drop detail, which is exactly the point.

Two to four spot inks

A spot ink is one solid colour printed on its own pass, the way a real stamp works. Limiting it to a handful of desaturated colours pulled from your own photo is what keeps the poster looking like it came from one place. The instruction not to force a fixed palette matters as much: a Jaipur photo and a Ladakh photo should not come back in the same three colours.

The printing defects

Dry ink gaps, notched contours, uneven pressure and one to two millimetres of misregistration are the tells that separate a carved stamp from clip art. Misregistration means the colour layers do not line up perfectly, because a hand pressing a stamp never lands twice in the same spot. Ask for clean edges and you get a logo.

The four lines of text

Place name, a record number, three keywords, a year. That is the entire copy deck, and the restriction is deliberate: field notes are written for the person who took them, not for an audience. The moment a slogan appears, the poster turns into an advertisement for a holiday.

Which photos give the best stamp

The prompt already tells the model how to carve each kind of scene, so you do not have to. Your only real decision is which photo to hand it. A good candidate has one clear silhouette you could still recognise if someone drew it from memory in five seconds.

• A single landmark with a strong outer shape: a dome, a fort wall, a spire, a stepwell.
• A hillside town where the houses stack in layers, which compresses beautifully into terraced blocks of ink.
• A coastline or a lake with a headland, where the shoreline itself is the drawing.
• A wide landscape with a distinct ridge, a lone tree, or a road running through it.

What struggles: interiors, close-ups of food, night shots where everything is silhouette anyway, and crowd photos with no structure behind them. There is nothing in those to carve, so the model invents something, and an invented stamp is obvious immediately.

What to change before you run it again

Run it once as written, then adjust one thing at a time. Changing three lines at once makes it impossible to tell which change helped.

• The text lines. If you want a specific place name, record number or year, write them into the prompt in quotes instead of letting the model choose. This also removes most spelling risk.
• The ink count. If the stamp comes back muddy, ask for two inks rather than four. Fewer colours almost always read as more deliberate.
• The stamp size. If it feels lost on the page, move the range up slightly, to about 38 to 42 per cent, and no further.
• The side the photo sits on. Flipping the layout so the paper falls on the left works, as long as you keep the proportions uneven.

Common mistakes when making rubber stamp travel posters in 2026

• Uploading several photos in one message. You get a collage, or one poster built from a blend of two places that exists nowhere.
• Asking for the stamp to include more of the scene. Every element you add costs you the recognisability of the ones already there.
• Filling the blank paper. The empty space is the design. Adding a border, a second stamp or a caption block turns a field note into a souvenir template.
• Accepting the first text output without reading it. A misspelt town name is the one flaw nobody reads as charming.
• Asking for it to look cleaner. Every request to tidy the edges pushes the result towards a vector logo, which is the one thing the prompt spends four lines trying to prevent.

FAQ

Does this work outside ChatGPT?

It works in any image tool that accepts a reference photo and follows long instructions, including Google's Gemini image model. Results vary in how strictly the photo is preserved, so check the left half carefully before you print. Tools that treat your upload as loose inspiration rather than source material are not worth the attempt.

Can I use it on photos of people?

You can, and the prompt tells the model to preserve identity, but portraits give the stamp very little to work with. A photo where a person stands inside a recognisable place will do far better than a close-up, because the place is what gets carved.

What size should I export for printing?

4x6 inches matches a standard postcard and suits the 4:3 frame closely enough that you lose almost nothing at the edges. For framing, print larger and ask for the highest resolution your tool offers, since the paper grain and ink texture are what carry the effect up close.

The photo came back subtly changed. What went wrong?

Usually the model got treated to too much freedom by an extra instruction, or the source image was low resolution and it filled in detail. Run it again with the preservation lines intact, and add a closing sentence of your own: keep the left photograph pixel-faithful to the upload, change nothing in it.

Can I make a set from one trip?

That is where it earns its keep. Run each photo separately, then number them in sequence when you write the text lines, so the record numbers run in order across the set. Six posters from one trip, printed at postcard size, read as a series rather than six versions of the same idea.

In short

This is a prompt about restraint rather than effects. It protects your photograph instead of restyling it, then earns the second half of the frame with one small carved mark and four lines of typing. Pick a photo with a shape worth remembering, run it on its own, read the text before you save it, and print at 4x6. Do that across a trip and you end up with a set of field notes, which is a better thing to keep than a folder of filtered pictures.