A painting to hang
The photo of the place that matters, in oil or watercolor, then enlarged by the Enhance node in faithful mode before printing. A canvas print of a scene no painter ever saw.
Turn a photo into a painting with AI
The same photo edited three times by Flux Kontext: impressionist oil, watercolor, pop art screen print. The scene does not change, only the medium. The full workflow, visible without an account.
Interactive demo
Real workflow, real results. Pan and zoom freely.
Use this workflowReal workflow: the same photo enters three Image nodes set to Flux Kontext, each with the prompt of one medium, and you compare.
At a glance
- What the workflow produces
- 3 images in 4:3
- Canvas structure
- 8 nodes wired by 6 connections
- Models wired in
- Flux Kontext
- Cost of one full run
- about 18 credits, roughly $0.22
A holiday photo becomes a painting you hang, if the painting keeps the photo. That is the whole difference between a filter and an edit: the filter lays a texture, the model repaints the scene. This workflow sends the same photo, a fishing harbor at dusk, into three nodes set to Flux Kontext, an editing model that changes the image it is given instead of inventing another one: impressionist oil, watercolor, pop art screen print.
Each prompt has two halves. The first describes the medium as a painter would: thick strokes and broken color for the oil, transparent washes and white paper for the watercolor, flat areas and halftone dots for the pop art. The second is the list of what does not move: the same boats in the same places, the same houses, the lighthouse, the framing, the evening light. Without that list the model paints another harbor, a pretty one, but it is no longer your photo.
The demo photo is generated, so that the page repaints nobody's holiday. Yours comes in through the Media node. The three paintings cost twenty-two cents, six credits each: less than a postcard.
How it works
- 1
Import a sharp, well-framed photo
A landscape, a street, a portrait, an animal. What is sharp becomes texture; what is blurred gets reinvented. A framing that is already composed gives a painting that is already composed.
- 2
Adapt the list of what does not move
"The same boats, the same lighthouse" describes the demo. Replace it with your scene: the people, the dog, the red car, the mountain. Everything that must stay in place.
- 3
Describe the medium, not a keyword
"Watercolor" alone gives anything. "Transparent washes, wet-on-wet bleeding in the sky, white paper in the highlights, pencil lines" gives a watercolor. Name a painter for a precise manner.
- 4
Compare the three, keep one
"Generate all" produces the three paintings. Look at the details you know: that is where a painting holds or betrays. Keep the one you like, or change one sentence and rerun that node alone.
Variations and use cases
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
A gift that does not exist in a shop
The family house, the dog, the first flat, in pop art for a kitchen or in watercolor for a living room. The painting tells a story the recipient knows.
An illustration for an article or a site
A bland stock photo becomes a unique watercolor, in the site's palette. The style is kept from one image to the next by copying the prompt, only the photo changes.
A painted portrait
The same mechanics on a face: the medium on one side, "same face, same expression, same pose" on the other. Pop art makes a studio portrait, oil makes a museum one.
A consistent series
Ten photos from a trip, one watercolor prompt copied into ten nodes: ten paintings that belong together. The travel sketchbook you wish you could have painted.
The painting, then animated
Once the painting is chosen, a Video node turns it into a five-second shot where the water ripples and the clouds drift. The animate an old photo workflow shows the principle.
Common mistakes
Forgetting the list of what does not move
"Make it a watercolor" without the lock gives a watercolor of another harbor. The second half of the prompt, the one that lists the objects, the places and the framing, is the one doing the work.
A keyword instead of a medium
"Painting style" describes nothing. Strokes, material, paper, halftone, palette: describe the medium as a painter. The three prompts of the demo are templates to copy.
Starting from a blurred or dark photo
The model repaints what it sees. What is blurred gets invented, what is crushed stays crushed. A sharp, lit, already well-framed photo gives the three best paintings.
Chaining two edits
Repainting the oil painting as pop art gives a pop art with the drift of both passes. Each medium starts from the original photo: one node per medium, all wired to the Media node.
Recommended models
Flux Kontext
Seedream 5 Pro
Topaz PrecisionFrequently asked questions
Does the scene really stay the same?
Yes when the list of what does not move is in the prompt, no when it is missing. In the demo the three paintings keep the four boats, the houses and the lighthouse in place.
Does it work on a portrait?
Yes, with the same list: "same face, same features, same expression, same pose". The model repaints the person without changing them. For a face as a cartoon rather than a painting, see photo to anime.
Can I ask for a specific painter?
A manner, yes: "in the manner of Monet", "in the spirit of Hokusai". The demo does it for the oil and the pop art. The model knows the great manners, not the works one by one.
How much does a full pass cost?
Eighteen credits for the three paintings, six per Flux Kontext render. The price shows on each node before you click. Rerunning a single medium costs only that one.
How big is the painting?
The format of your photo, 4:3 in the demo, around two megapixels. To print large, the Enhance node in faithful mode takes it four times bigger without reinventing the strokes.
Can I do it from a phone?
Yes, and it is the natural use: a photo from the gallery, imported into the Media node on the mobile canvas. The create on mobile guide explains the gestures.
Which other mediums work?
Pastel, charcoal, etching, Indian ink, gouache, mosaic, stained glass, pixel art. Describe the material and the mark of the tool, and keep the list of what does not move. One node per medium.
Are my photos used to train the models?
No. What you import stays in your workspace and is passed to no training. The demo photo is generated for the same reason.
Turn a photo into a painting with AI
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.