MuseFrame — AI Photo Art in 30 Styles

One photo, two directions

The source is a bus on an ordinary street. Drag the slider to compare it with Quiet Frame. The print beside it went the Zine Poster way.

The framing, the cars and the tree shadows stay where they were. What changes is colour, texture and light. We blurred the plates for this page; the generation leaves them alone.

Result: Quiet Frame style, a muted teal cinematic grade over the street and bus — **Original / Quiet Frame**Drag across the image to compare

Original photo: a city street under green trees with a red-and-white bus and two white cars — **Original / Quiet Frame**Drag across the image to compare

The same street photo as a Zine Poster: torn black-and-white halftone fragments on scanned cream paper, the bus sign kept in red, sparse typewriter notes around it — **Zine Poster**Small Press · Exhibition 07Torn halftone, with the bus sign left red.

Portrait before-and-afters need the sitter’s consent, and we’re still collecting them. Meanwhile, how likeness is protected and which photos tend to fail is covered in Why AI portraits stop looking like you.

Seven exhibitions, thirty styles

The first six exhibitions hold 24 styles we made ourselves. The seventh, Small Press, adds six adapted from community work.

Each style is a versioned spec that says what it’s for and what it must not change. None of them imitates a living artist or a brand.

Edition 02 · Small Press

Small Press

Zine Poster, Cover Story, Reportage Wash, Ink & Seal, One Line and Studio Hour. They’re adapted from MIT-licensed community skills, and each spec records the source and licence.

Designed text is allowed into the frame here. Real brands and logos never are.

View all 6

All 30 styles

The same list you’ll find inside each exhibition. Styles marked Premium need a Creator subscription.

Exhibition 01 · 4 stylesQuiet Portraits

Soft WindowNear-window light, gentle contrast.Portraits
Muted RoomQuiet interiors, clay and dust tones.Portraits, objects
Silver AirPremiumCool silver cast, fine grain.Portraits
Still GazeHeld expression, deep neutral ground.Portraits

Exhibition 02 · 4 stylesPrinted Matter

Ink RegisterOffset inks on uncoated paper.Portraits, objects
Paper CutFlat shapes with a clean blade edge.Objects, portraits
Dry PressChalky blacks, pressed texture.Portraits
Sunday InsertPremiumHalftone warmth, weekend colour.Portraits, objects

Exhibition 03 · 4 stylesDream Geography

Mist AtlasFog layers over slow water.Landscapes
Night OrchardPremiumBlue night, lantern warmth.Landscapes, portraits
Blue DistanceHaze that deepens with distance.Landscapes, portraits
Floating WeatherLifted pastel skies, soft mass.Landscapes

Exhibition 04 · 4 stylesGraphic Light

Cobalt ShadowOne blue, cut by hard shadow.Portraits, architecture
Hard NoonNoon contrast, bleached ground.Architecture, landscapes
Color BlockPremiumThree fields, one subject.Portraits
AfterimageChromatic echo at the edges.Portraits

Exhibition 05 · 4 stylesMaterial Studies

Woven LightFabric weave carrying warm light.Objects, pets, portraits
Mineral WashStone pigment, cool sediment.Objects, landscapes
Wax SurfaceSealed glow, softened edges.Objects, portraits
Charcoal BloomAsh blacks with a breathing core.Portraits, objects

Exhibition 06 · 4 stylesSmall Cinemas

Quiet FrameMuted teal air, patient framing.Portraits, street
Late PlatformSodium lamps, last-train tension.Street, portraits
Warm InteriorTungsten rooms, honest faces.Portraits
Distant FlashPremiumCold flash, caught mid-thought.Portraits, street

Exhibition 07 · 6 stylesSmall Press

Zine PosterTorn paper, one loud colour, quiet notes.Portraits, landscapes, objects
Cover StoryYou, on this month’s witty cover.Portraits, pets
Reportage WashInk lines, watercolour air, field notes.Portraits, street, objects
Ink & SealFive tones of ink, one red seal.Landscapes, objects, portraits
One LineOne continuous line, one quiet accent.Portraits, objects
Studio HourPremiumSoft-box light, your best hour.Portraits

Four steps, about two minutes

One thing per screen. The progress screen shows the four real stages, never a made-up percentage.

I
Pick a photo
Choose one from your library or take one with your phone. It’s re-encoded as a JPEG on your device, with GPS and camera details removed, before anything uploads.
In the browser, the browser does the re-encoding.
II
Read the image
The app checks the subject, how many people there are, the exposure and the sharpness. Then it suggests styles and says why, such as “preserves soft face light”.
III
Choose a direction
Pick a style and you have three calls left: how strong the style is, how closely the subject follows the original, and whether to keep or reframe the composition.
IV
Generate and save
It works through four stages: preparing, building the direction, making the image, checking it. Hold to see the original, drag to compare, then save or share.

Portraits, group shots and still life hold up best. Wide landscapes are not its strength.

Your photo, used once

We don’t train models on your photos or build face databases. We don’t sell your data to anyone.

Payments go through the app store or our payment processor. We never store your card.

Location gone before upload: Your photo is re-encoded on your device, so GPS and camera details are removed before the file leaves your phone.
Server deletion is done by hand: Deleting a project removes it from your library straight away. To clear the original and the result from our servers as well, email us with the project and we’ll reply once it’s done. Closing your account works the same way; see Account deletion.
Hold to see the original: Every result flips back to the original with one press, and a slider compares them region by region. You can see for yourself what changed.

About AI-generated content

Final images are currently generated or edited by OpenAI GPT Image 2 (gpt-image-2). Some designed styles use OpenAI GPT-5.4 mini (gpt-5.4-mini) to compile style instructions; it does not generate the final image. If the model or provider changes for safety, availability or product reasons, we update the Terms, Privacy Policy and Acceptable Use Policy.

Three free, then pay as you go

Sign up with your email and you get 3 complete images, with no password and no card. After that, one credit makes one standard image, and a failed or rejected generation costs nothing.

Packs never expire and give you the standard tier with every non-Premium style. Premium styles and the high tier come with Creator.

Free on sign-up
Email code, no password, no card. No watermark, every non-Premium style.
Free
3 images
Sign up · 3 free
Trial 3
Never expires. One per account.
$1.99one-time
$0.66 per image
Buy 3
Pack 10
Never expires. Top up when you need to.
$5.99one-time
$0.60 per image
Buy 10
Pack 30Best value
Never expires. Top up when you need to.
$12.99one-time
$0.43 per image
Buy 30
Pack 100
Never expires. Top up when you need to.
$34.99one-time
$0.35 per image
Buy 100
Creator Monthly
Every Premium style and the high tier. 30 images a month, valid for that month. Cancel any time.
$7.99/ month
$0.27 per image
Subscribe
Creator Yearly
Every Premium style and the high tier. 360 images up front, valid for the whole year.
$59.99/ year
about $5 a month
Subscribe

These USD prices are for the English site and app. Chinese-language users see a separate CNY list, with a 30-day Creator pass in place of a subscription. The two lists are set separately, not converted. Web subscriptions are billed in USD only.

How to pay

In the web app, pick a plan and you’ll go to a checkout page run by our payment processor, Waffo Pancake. Card, Apple Pay and Google Pay all work, and the credits arrive as soon as the payment clears.

You can also join QQ group 824558022 and message the admin with your sign-up email and the plan you want, or write to donaldkuke@gmail.com. We top up by hand, usually within minutes during the day.

Questions people ask

Do I need to write a prompt?

No. You pick one of 30 styles across 7 exhibitions, then set style strength, subject fidelity and composition. The style’s spec takes care of the rest.

Will it still look like me?

That’s a fixed rule in every style. Facial structure, expression, pose, hairstyle, skin tone and accessories like glasses stay as they were. Nobody gets prettified or swapped, the number of people stays the same, and nothing new is added to the scene.

It can still miss now and then, more often with very poor light or small faces. On the result screen, hold the image to see the original.

How long does one artwork take?

Usually about two minutes, and 1 to 5 minutes at busy times. You can leave while it runs and find the result under Projects later.

What happens to my photo?

It’s re-encoded on your device, which strips GPS and camera details before upload. We use it only to make the image you asked for. We don’t train models on it, build face databases with it or sell it.

Deleting a project removes it from your library. If you want the original and the result cleared from our servers as well, email us. We do it by hand and reply when it’s done.

Is MuseFrame free?

Sign up by email and your first 3 images are free, complete and without a watermark. After that you can buy a pack: Trial 3 for $1.99 (one per account), 10 for $5.99, 30 for $12.99 or 100 for $34.99. Packs never expire.

For Premium styles and the high tier, Creator is $7.99 a month with 30 images, or $59.99 a year with 360 images granted up front and valid for the year.

Pay inside the web app by card, Apple Pay or Google Pay, or ask in QQ group 824558022. Failed or safety-rejected generations never use a credit.

Which platforms are supported?

The web app is live and works in a phone browser. The Android app is finished and opens for download soon. iOS is planned.