Travel Milk Frog (旅行奶蛙) — Pack a Bag, Send It Out, Bring Home a Masterpiece | Neta Studio
Play Travel Milk Frog (旅行奶蛙), a free browser pixel-cabin collecting game where you harvest money grass, pack the butter-yellow milk-frog mascot a bag and send it out to bring back 94 milk-frog reworkings of real world masterpieces across 16 galleries — plus letters and travel postcards for the wall. No install, no signup, no ads. Built and published in Neta Studio.
Final Product: Travel Milk Frog (旅行奶蛙)
Cover
The page key art, generated 16:9 with GPT Image 2 from the world’s own character lock: the canonical butter-yellow milk frog, bow tie on, holds a folded map and a leather satchel at the gate of its red-roofed cottage, while milk-frog reworkings of The Great Wave, The Starry Night and the Mona Lisa float in gilded frames above a clover meadow, with a handwritten letter and a postcard drifting past.
Screenshot
The real game, opening view. The 320×180 pixel scene renders a red-roofed cabin, picket fence and a sandy path where the frog waits, with a live stat bar (money grass, collected works out of 94, unread letters, trips taken) and the five tabs 旅行角 · 信箱 · 明信片墙 · 图鉴 · 记录 underneath. The hint on the meadow reads “点草地上的钱草收一收 · 按住扫过收更多” — tap the money grass, or press and sweep to harvest much more.
Screenshot
The heart of the loop: three numbered choices. ① 挑一件兴趣物 picks one of five interest items that steers which eras the frog gravitates to, ② 塞一份补给 packs one single-use supply, and ③ 选好时长 sets how long the trip lasts. Every combination changes both the loot that comes back and the story it tells.
Travel Milk Frog (旅行奶蛙) is a free browser pixel-cabin collecting game generated by Neta Studio inside the milk-frog museum world. You keep a sunny low-resolution cottage, harvest money grass from the meadow, then pack the milk-frog mascot a bag: one interest item that steers which eras it gravitates to, one single-use supply, and a trip length. Send it out and wait — while it is away it walks real galleries and comes home with a milk-frog reworking of a famous painting, a letter in its own voice and a few travel photographs that go straight up on the postcard wall, all filed into an album that fills toward 94 works across 16 galleries.
It is one static pixel-art web app: no install, no account, no ads and no purchases. Everything is drawn at the game’s native 320×180 pixel resolution and integer-scaled, so the whole thing loads in about a second on any device. The frog’s collection overlaps the other milk-frog sets — the museum, the card pack and the stamp album — but the routes, the rarities and the framing are its own.
Play it now: Travel Milk Frog →
Features
- 94 Real Masterpieces, 16 Galleries — The catalogue holds 94 milk-frog reworkings of real paintings, spanning 16 galleries from prehistory and Qin–Han China through the Renaissance, Romanticism, Ukiyo-e, Impressionism and Modernism. Every collected card names the original work and artist and links straight to the museum entry — so the album you build is a real art-history syllabus, not filler.
- Three Choices Plan Every Trip — Each journey is three picks: one of five interest items (a folding map, an oil brush, a ukiyo-e block, a telescope or a small camera) that decides which eras you are most likely to meet, one of five single-use supplies (rice ball, dried fish, picnic box, milk tablet or blank postcard) and one of four trip lengths, from a free thirty-minute stroll to a full-day roam that first needs twelve works in the album.
- A Postcard Wall That Fills Up — Every trip the frog takes brings back travel photos, and each one goes up on the cottage’s 明信片墙. Leaf through them one at a time: each postcard carries two layers — a milk frog inside a real masterpiece, plus a short note from the road such as “This painting is bigger than I expected. You have to look up to see all of it.”
- Letters in the Frog’s Own Voice — Trips do not just produce art. The 信箱 fills with first-person letters that say how long it walked and how many rooms it entered, and the 图鉴 and 记录 tabs keep the full album and a running log of every outing — so the world accumulates a diary as well as a collection.
- Money Grass, Not Microtransactions — The only currency is money grass growing in front of the cabin: tap a clump to collect it, or press and sweep across the meadow to gather far more at once. It pays for interest items, supplies and longer trips. There are no ads, no in-app purchases and no energy timers to pay away.
- True 320×180 Pixel Craft — The cabin scene, the travelling frog, the harvest animation and every icon are drawn in code at a 320×180 buffer and integer-scaled, so it stays crisp on phones and desktops alike. The whole game is one small static bundle that opens in about a second and saves your coins, album, letters and postcards in the browser.
Interface
UI
The 图鉴 album is the collection itself: a grid of milk-frog masterpieces, each card naming the work, the era and the original artist — here a Kandinsky, a Monet water-lily pond, Hokusai’s Great Wave, Munch’s Scream and Van Gogh’s Starry Night, all repainted with the frog as the sitter and linked to the museum original.
UI
The 明信片墙 hangs the travel photographs the frog brings home. Each postcard pairs the artwork it saw with a note from the road — “The road bent and suddenly opened up”, “There’s nobody in the gallery, only this one painting is lit” — and links through to the full museum entry.
UI
The 信箱 keeps the letters. Each one is written in the frog’s own voice: how many minutes it walked, how many rooms it entered, which painting it stood in front of the longest — and a direct link to go and see that work in the museum.
Sprite
One of the character sprites cut to a transparent PNG with BiRefNet. The official butter-yellow milk frog — smooth matte body that flows head-first into the torso with no neck, pale cream belly, two pale-green eye rings with centred black pupils, a tiny mouth and short taupe mitten hands.
Sprite
The same identity at rest, cut out cleanly so it drops onto any background. Because the character lock is fixed, every sprite, painting and key-art frame stays recognisably the same frog.
Sprite
A three-quarter sprite from the same reference set. The cut-outs feed the game’s poses and the marketing page without a re-render.
Sprite
The travelling pose, isolated on transparency with BiRefNet-general and saved as a real PNG with an alpha channel.
Production Process Breakdown
Stage 1: World and Character Designclaude-opus-5-5
The world begins from a single line — “I want to build a milk-frog museum site that exhibits world-famous paintings and artworks reworked with milk frogs; the UI and layout should reference breguet.com, very high-end.” From that, the canon is locked: one official Milk Frog character atom fixes the pear-shaped body that flows into the torso with no neck, the large pale-cream belly, two pale-green eye rings each holding a flat black pupil, one tiny straight mouth line and short taupe mitten hands. Travel Milk Frog keeps that identity untouched and re-frames it as a cosy pixel-cabin collecting loop instead of a museum page.
The world begins from a single line — “I want to build a milk-frog museum site that exhibits world-famous paintings and artworks reworked with milk frogs; the UI and layout should reference breguet.com, very high-end.” From that, the canon is locked: one official Milk Frog character atom fixes the pear-shaped body that flows into the torso with no neck, the large pale-cream belly, two pale-green eye rings each holding a flat black pupil, one tiny straight mouth line and short taupe mitten hands. Travel Milk Frog keeps that identity untouched and re-frames it as a cosy pixel-cabin collecting loop instead of a museum page.
Stage 2: Asset Direction and Generationgpt-image-2birefnet-general
Art direction is fixed before anything is drawn. The 94 painting adaptations and the official Milk Frog reference sheets are generated with GPT Image 2, the 16:9 key-art cover is generated from the same identity lock, and the character art is cut to transparent PNG sprites with BiRefNet. The game scene itself is drawn in code at a 320×180 pixel buffer so it stays crisp when integer-scaled.
Art direction is fixed before anything is drawn. The 94 painting adaptations and the official Milk Frog reference sheets are generated with GPT Image 2, the 16:9 key-art cover is generated from the same identity lock, and the character art is cut to transparent PNG sprites with BiRefNet. The game scene itself is drawn in code at a 320×180 pixel buffer so it stays crisp when integer-scaled.
Stage 3: Build, QA and Publishdeepseek-v4-pro
The game ships as a small static web app — one HTML document, one stylesheet and a handful of scripts driving a canvas pixel engine — so it loads instantly and needs no backend. QA opens it in a real browser, sweeps the money-grass meadow, runs every interest item, supply and trip length through a full loop, checks that each collected card links to the right museum original, and then publishes the finished work to a shareable browser URL.
The game ships as a small static web app — one HTML document, one stylesheet and a handful of scripts driving a canvas pixel engine — so it loads instantly and needs no backend. QA opens it in a real browser, sweeps the money-grass meadow, runs every interest item, supply and trip length through a full loop, checks that each collected card links to the right museum original, and then publishes the finished work to a shareable browser URL.
Asset Prompts
Real Generation Prompts
Mona Lisa as a Milk-Frog Oil Painting Promptgpt-image-2
“Reimagine Leonardo da Vinci's Mona Lisa as an oil painting, but the sitter is the round butter-yellow milk-frog character from the reference (soft plump body, cream-colored belly, calm green eyes, gentle closed-mouth smile). The frog sits in the exact three-quarter Mona Lisa pose, hands folded, before the hazy Renaissance mountain landscape with winding river. Sfumato technique, cracked aged varnish, dark museum-oil palette, golden-brown tones. Faithful to the original composition and lighting. No text.”
“Reimagine Leonardo da Vinci's Mona Lisa as an oil painting, but the sitter is the round butter-yellow milk-frog character from the reference (soft plump body, cream-colored belly, calm green eyes, gentle closed-mouth smile). The frog sits in the exact three-quarter Mona Lisa pose, hands folded, before the hazy Renaissance mountain landscape with winding river. Sfumato technique, cracked aged varnish, dark museum-oil palette, golden-brown tones. Faithful to the original composition and lighting. No text.”
Botticelli’s The Birth of Venus as a Milk-Frog Painting Promptgpt-image-2
“Reimagine Botticelli's The Birth of Venus. A plump butter-yellow milk-frog stands demurely on a giant floating scallop shell at the center, cream belly forward, stubby arms in the modest Venus pose, calm round eyes. To the left two frog wind-gods blow breezes; to the right a frog attendant offers a floral cloak. Pale tempera palette, soft greens and gold, gentle sea and sky, Early-Renaissance line and delicacy. CRITICAL RENDERING RULE: Use the stylized simple cartoon milk-frog from the reference image — smooth matte butter-yellow rounded body, large cream/ivory belly, simple round eyes, tiny calm mouth, stubby grey-tipped limbs. NOT a realistic textured frog. NOT a frog head grafted onto a human body. The plump frog IS the entire figure, repainted completely in the target masterpiece style so it dissolves into the original brushwork, texture and palette. No text, no signature, no watermark, no border captions.”
“Reimagine Botticelli's The Birth of Venus. A plump butter-yellow milk-frog stands demurely on a giant floating scallop shell at the center, cream belly forward, stubby arms in the modest Venus pose, calm round eyes. To the left two frog wind-gods blow breezes; to the right a frog attendant offers a floral cloak. Pale tempera palette, soft greens and gold, gentle sea and sky, Early-Renaissance line and delicacy. CRITICAL RENDERING RULE: Use the stylized simple cartoon milk-frog from the reference image — smooth matte butter-yellow rounded body, large cream/ivory belly, simple round eyes, tiny calm mouth, stubby grey-tipped limbs. NOT a realistic textured frog. NOT a frog head grafted onto a human body. The plump frog IS the entire figure, repainted completely in the target masterpiece style so it dissolves into the original brushwork, texture and palette. No text, no signature, no watermark, no border captions.”
Vermeer’s The Milkmaid as a Milk-Frog Painting Promptgpt-image-2
“Reimagine Vermeer's The Milkmaid. A sturdy milk-frog in a linen cap, yellow apron and blue skirt stands at a table pouring a thin stream of milk from an earthenware jug into a bowl, cream belly rounded, calm eyes down at its work. Soft daylight from a left window, bread and crockery on the table, plaster wall, quiet Dutch domestic oil realism. CRITICAL RENDERING RULE: Use the stylized simple cartoon milk-frog from the reference image — smooth matte butter-yellow rounded body, large cream/ivory belly, simple round eyes, tiny calm mouth, stubby grey-tipped limbs. NOT a realistic textured frog. NOT a frog head grafted onto a human body. The plump frog IS the entire figure, repainted completely in the target masterpiece style so it dissolves into the original brushwork, texture and palette. No text, no signature, no watermark, no border captions.”
“Reimagine Vermeer's The Milkmaid. A sturdy milk-frog in a linen cap, yellow apron and blue skirt stands at a table pouring a thin stream of milk from an earthenware jug into a bowl, cream belly rounded, calm eyes down at its work. Soft daylight from a left window, bread and crockery on the table, plaster wall, quiet Dutch domestic oil realism. CRITICAL RENDERING RULE: Use the stylized simple cartoon milk-frog from the reference image — smooth matte butter-yellow rounded body, large cream/ivory belly, simple round eyes, tiny calm mouth, stubby grey-tipped limbs. NOT a realistic textured frog. NOT a frog head grafted onto a human body. The plump frog IS the entire figure, repainted completely in the target masterpiece style so it dissolves into the original brushwork, texture and palette. No text, no signature, no watermark, no border captions.”
Vermeer’s Girl with a Pearl Earring as a Milk-Frog Painting Promptgpt-image-2
“Create an edge-to-edge museum-quality oil painting closely recreating Johannes Vermeer's Girl with a Pearl Earring, with the attached milk-frog character replacing the girl as the sole sitter. Study the supplied correct example: the milk-frog is captured by preserving its original simple character silhouette, not by inventing frog anatomy. Character fidelity: smooth matte golden-yellow rounded head and upper body, no visible nose, no nostrils, no muzzle, no cheek modeling, no lips; only two wide-set circular eyes with pale green iris rings and black pupils, plus one tiny thin horizontal mouth line. The face is simple, calm and nearly flat, with gentle painterly shading only. The milk-frog turns three-quarter toward the viewer over its shoulder exactly like Vermeer's sitter, with the rounded yellow shoulder and torso continuing naturally below the head. It wears Vermeer's blue and ochre headscarf and ochre-brown jacket, and one luminous pearl earring hangs beside the face. Use Vermeer's dark near-black background, soft single window light from the left, quiet blue-and-gold palette, delicate oil glazing and subtle old-canvas texture. The milk-frog must feel painted into Vermeer's work, while retaining the distinctive simple reference face. No exposed full belly, no full-body pose, no human anatomy, no frog nostrils, no realistic frog skin, no giant mouth, no teeth, no eyebrows, no hair, no long human fingers. Absolutely no frame, plaque, caption, title, letters, numbers, signature, watermark, typography or border.”
“Create an edge-to-edge museum-quality oil painting closely recreating Johannes Vermeer's Girl with a Pearl Earring, with the attached milk-frog character replacing the girl as the sole sitter. Study the supplied correct example: the milk-frog is captured by preserving its original simple character silhouette, not by inventing frog anatomy. Character fidelity: smooth matte golden-yellow rounded head and upper body, no visible nose, no nostrils, no muzzle, no cheek modeling, no lips; only two wide-set circular eyes with pale green iris rings and black pupils, plus one tiny thin horizontal mouth line. The face is simple, calm and nearly flat, with gentle painterly shading only. The milk-frog turns three-quarter toward the viewer over its shoulder exactly like Vermeer's sitter, with the rounded yellow shoulder and torso continuing naturally below the head. It wears Vermeer's blue and ochre headscarf and ochre-brown jacket, and one luminous pearl earring hangs beside the face. Use Vermeer's dark near-black background, soft single window light from the left, quiet blue-and-gold palette, delicate oil glazing and subtle old-canvas texture. The milk-frog must feel painted into Vermeer's work, while retaining the distinctive simple reference face. No exposed full belly, no full-body pose, no human anatomy, no frog nostrils, no realistic frog skin, no giant mouth, no teeth, no eyebrows, no hair, no long human fingers. Absolutely no frame, plaque, caption, title, letters, numbers, signature, watermark, typography or border.”
Edge-to-Edge Renaissance Oil Rework Promptgpt-image-2
“Create an edge-to-edge unframed Renaissance oil painting closely matching the original composition, pose, crop and quiet mood of Leonardo da Vinci's Mona Lisa. The sitter is Mona Lisa reimagined as the milk-frog character, but preserve the original portrait's human-like seated presentation and conceal most of the body beneath the original dark Renaissance gown. Show only the recognizable milk-frog traits: smooth matte golden-yellow skin, a softly rounded simplified frog-like face with no nose and absolutely no nostrils, two wide-set protruding eyes with pale green irises and black pupils, and a very small thin horizontal mouth. Keep the face flatter and simpler, like the original stylized 3D character reference, not a realistic amphibian and not a highly modeled frog head. Do not show a huge exposed belly or full frog anatomy; the dark gown should cover the torso and lower body as in the original painting, with only a modest cream-colored chest area suggested beneath the neckline. Preserve Mona Lisa's three-quarter turn, subtle calm expression, folded hands, dark sleeves, centered half-length portrait scale, hazy blue-green mountain landscape, winding river and distant bridge. Translate the yellow milk-frog face into Leonardo's soft sfumato and muted aged oil palette while keeping the simple character geometry. The result should feel like the original Mona Lisa with the milk-frog naturally painted into it, not a frog portrait pasted over it. ABSOLUTELY NO nostrils, nose, realistic frog skin, exaggerated amphibian mouth, giant exposed belly, full-body pose, picture frame, mat, plaque, museum label, caption, title, letters, numbers, signature, watermark, typography or border.”
“Create an edge-to-edge unframed Renaissance oil painting closely matching the original composition, pose, crop and quiet mood of Leonardo da Vinci's Mona Lisa. The sitter is Mona Lisa reimagined as the milk-frog character, but preserve the original portrait's human-like seated presentation and conceal most of the body beneath the original dark Renaissance gown. Show only the recognizable milk-frog traits: smooth matte golden-yellow skin, a softly rounded simplified frog-like face with no nose and absolutely no nostrils, two wide-set protruding eyes with pale green irises and black pupils, and a very small thin horizontal mouth. Keep the face flatter and simpler, like the original stylized 3D character reference, not a realistic amphibian and not a highly modeled frog head. Do not show a huge exposed belly or full frog anatomy; the dark gown should cover the torso and lower body as in the original painting, with only a modest cream-colored chest area suggested beneath the neckline. Preserve Mona Lisa's three-quarter turn, subtle calm expression, folded hands, dark sleeves, centered half-length portrait scale, hazy blue-green mountain landscape, winding river and distant bridge. Translate the yellow milk-frog face into Leonardo's soft sfumato and muted aged oil palette while keeping the simple character geometry. The result should feel like the original Mona Lisa with the milk-frog naturally painted into it, not a frog portrait pasted over it. ABSOLUTELY NO nostrils, nose, realistic frog skin, exaggerated amphibian mouth, giant exposed belly, full-body pose, picture frame, mat, plaque, museum label, caption, title, letters, numbers, signature, watermark, typography or border.”
Official Milk Frog Default Avatar Promptgpt-image-2
“Reference 1 is a default user profile placeholder avatar: a flat light-gray circle head above a flat light-gray semicircle torso, centered on a solid medium-gray square background. Reference 2 is the official character sheet of a mascot creature: a plump, round, blob-shaped yellow frog-like character whose head and body merge into one smooth chubby dome, with LARGE round green eyes (soft green iris, black pupil) set wide apart facing FORWARD on the front of its face, a tiny straight neutral mouth, and a pale belly. Redo reference 1 as an avatar of THIS character, keeping reference 1's minimalist flat style and gray palette almost exactly: one plump rounded blob-shaped head-and-body silhouette (head and torso merged into a single chubby dome, much rounder and plumper than the original human head circle), flat-cropped at the bottom edge like the original bust, rendered in the same flat light gray on the same solid medium-gray background. Give her the character's signature LARGE round green eyes with black pupils placed wide apart on the front of the face (the only color accent, soft muted green), and a tiny straight neutral mouth. NO bulging eyeball bumps on top of the head, no regular-frog look, no smile curve, no arms, no belly patch, no gradients, no outlines, no shadows, no texture, no text. Flat solid fills, square profile avatar, centered, clean and minimal.”
“Reference 1 is a default user profile placeholder avatar: a flat light-gray circle head above a flat light-gray semicircle torso, centered on a solid medium-gray square background. Reference 2 is the official character sheet of a mascot creature: a plump, round, blob-shaped yellow frog-like character whose head and body merge into one smooth chubby dome, with LARGE round green eyes (soft green iris, black pupil) set wide apart facing FORWARD on the front of its face, a tiny straight neutral mouth, and a pale belly. Redo reference 1 as an avatar of THIS character, keeping reference 1's minimalist flat style and gray palette almost exactly: one plump rounded blob-shaped head-and-body silhouette (head and torso merged into a single chubby dome, much rounder and plumper than the original human head circle), flat-cropped at the bottom edge like the original bust, rendered in the same flat light gray on the same solid medium-gray background. Give her the character's signature LARGE round green eyes with black pupils placed wide apart on the front of the face (the only color accent, soft muted green), and a tiny straight neutral mouth. NO bulging eyeball bumps on top of the head, no regular-frog look, no smile curve, no arms, no belly patch, no gradients, no outlines, no shadows, no texture, no text. Flat solid fills, square profile avatar, centered, clean and minimal.”
Why Choose Neta Studio
Professional — The cloud harness supports the full pipeline from world building to a finished, shareable interactive game, all running in the cloud with no local setup and stable access to the latest models: Claude Opus 5, GPT Image 2, BiRefNet and DeepSeek V4.
One-click generation — Officially tuned skills drive every step: world building, character design, art generation, pixel-art layout and publishing. World forking lets you inherit characters, locations and rules from any public world, then rewrite the parts you want.
Community-first — Users build unlimited worlds, and every public world can be forked, remixed and reimagined. The Discord learning community provides tutorials and real-time support — from your first world to your first playable game.
Studio Interior
‹
›
Studio
Studio
Studio
Play It Now
This game was generated by Neta Studio AI inside the milk-frog museum world from a single seed prompt. The world took minutes. The game took minutes. You can do it too — fork this world or start fresh from any idea. No coding required, free to start.
×