All products

Product 01 / 03 Available

Un-Plastic

Turn waxy AI skin into real skin, in the photo you already made, without changing the face.

Before and after: an AI skincare portrait with waxy, plastic skin, fixed with the Un-Plastic Nano Banana Pro edit prompt. Same face and crop, now with visible pores.

Drag to compare

Make AI portraits look like real photos, without changing the face

AI image models smooth skin into wax: no pores, one flat tone, pure-white eyes and a highlight that looks painted on. Un-Plastic is two paste-in edit prompts that fix exactly that in the photo you already made. Attach it in Nano Banana Pro or GPT Image (ChatGPT) edit mode, paste one block, and the same photo comes back with real skin texture: visible pores, a natural sebum sheen, under-eye texture, fine facial hair and eyes with a real catchlight.

Everything else stays put. The face, expression, gaze, pose, hands, clothing, background, crop and lighting are locked in the prompt, so it works on finished client images, UGC ads and AI influencer posts without starting over.

Version 2.0 adds a second step. Run the edited photo through the Magnific Precision V2 upscaler at the settings in the guide and the whole image turns crisp and photographic. The guide names where to run it and what it costs: about $0.12 an image on FLORA, paid to FLORA. Un-Plastic does not include tool credits.

Will an upscaler fix plastic skin?

Upscaling is the obvious fix, and so is adding “8K” to the prompt; every before here was made with “8K” in the prompt. On all three test photos, Magnific Precision V2 on its own made the skin sharper but added no pores. The block adds the texture and the upscale adds the crispness: block first, then upscale.

Close-up of an AI skincare portrait: glossy, poreless skin on the nose and cheeks.
The AI photo
The same close-up after a Magnific Precision V2 upscale alone: sharper, but the skin is still poreless.
Upscale only
The same close-up after Block A in Nano Banana Pro: visible pores across the nose and cheeks.
Block A only
The same close-up after Block A and then the upscale: pores and freckles, and the whole image crisper.
Block A, then the upscale

Close-ups of test photo 10, the skincare ad, all four scaled to the same size. The upscale is Magnific Precision V2 at Set 1, run on FLORA.

The 2.0 recipe, up close

Drag to compare. The before is a plain “flawless, 8K” render from Nano Banana Pro at 4K. The after is the same photo after one run of Block A and one Magnific Precision V2 upscale at Set 1.

Close-up before and after of an AI UGC skincare photo: waxy skin fixed into visible pores and natural redness with the Un-Plastic prompt and a Magnific Precision V2 upscale, same smile and jar.
UGC skincare shot. Pores and a little redness across the nose and cheeks; the same smile, eyes and jar.
Close-up before and after of an AI LinkedIn headshot of a man in his 40s: airbrushed skin fixed into open pores, laugh lines and stubble, same face and smile.
LinkedIn headshot. Open pores, laugh lines and stubble come back. He still looks like him.
Close-up before and after of an AI portrait of an older Indian man with a grey beard: smoothed skin fixed into deep natural wrinkles and texture, same smile and warm light.
Older face, deep skin. Every wrinkle kept and sharper, and the warm street light stays warm.

Close-ups from six launch photos made on Nano Banana Pro at 4K; both sides are shown at the same size. All six kept the face, the smile and the gaze.

Before and after: ten waxy portraits, tested

Drag any image to compare. Each after is a single run of Block A in Nano Banana Pro, with nothing added, on a portrait generated to look as plastic as possible.

10/10
faces, smiles and gazes held in Nano Banana Pro at 4K
10/10
faces held in GPT Image edit mode at 1K, with a gentler skin change
4
rounds of testing on the two blocks
Before and after: AI portrait of a Black woman with deep skin, plastic skin fixed into natural texture with the same smile and warm skin tone.
Deep skin. Same smile and teeth. The glow is sheen along the cheekbones, the brown stays warm, and it is never ashy.
Before and after: AI portrait of a 60-year-old woman, waxy skin fixed while keeping her natural wrinkles.
Older face. The lines around the eyes and forehead are kept, not smoothed away.
Before and after: AI cafe selfie of an East Asian woman, plastic skin fixed with the same smile and framing.
Candid selfie. The smile and teeth stay exactly as they were; pores and faint freckles come through.
Before and after: AI portrait of a South Asian woman holding chai, skin fixed while her downcast gaze and golden light stay the same.
Eyes looking down. The eyes stay lowered to the cup, and the warm evening light is untouched.
Before and after: close-up AI portrait of a bearded man, airbrushed skin fixed into realistic pores and stubble.
Beard, close-up. Beard edges and expression held, with real pore and stubble texture.
Before and after: AI beauty portrait of a woman holding a serum bottle, plastic skin fixed while the product and hand stay unchanged.
Product in hand. Hand, bottle and label untouched; pores and faint freckles across the nose and cheeks.

How it was tested: ten portraits were generated from plain “flawless, 8K, studio lighting” prompts, across deep and light skin, an older face, men and women, glasses, phone selfies, half-body shots and products in hand. Each was run once through each block. A pair passed only if it came back as the same photo, with the same face, expression, gaze, pose and crop, and with real texture where the plastic had been.

How it works

  1. Open Nano Banana Pro or GPT Image (including ChatGPT) in edit mode and attach the AI photo that looks plastic.
  2. Paste the block for that model, exactly as written.
  3. Generate, then check that the face and the crop held. If your tool asks for a size, match your photo.
  4. Upscale the result with Magnific Precision V2 at the settings in the guide, about $0.12 an image on FLORA. This step is optional.
  5. Still waxy? Try the one-line helper. Face drifted? Try the shorter block. Both are included.

Who it’s for

Product-stills freelancers and small brands whose last delivery came back with “can you make it less AI?”

Inside

  • Block A for Nano Banana Pro edit mode: brings back pores, sheen and real eyes without redrawing the face
  • Block B for GPT Image and ChatGPT edit mode: a gentler fix, for when the face must not move at all
  • Tested Magnific Precision V2 settings, because in testing the upscale on its own added no pores: one set for real photos, one for clean ads
  • Two helper lines, for a photo that is still waxy or a face that drifts
  • A 12-page PDF guide with tested before-and-after pairs, so you know what a good result looks like
  • A plain text file, so the blocks paste in clean

How you get it

Pay once through a hosted checkout. The download link lands in your email within a minute, and the files are yours to keep.

Questions

Why do AI portraits look plastic?

Image models average away the small things that make skin read as real: pores, slight changes in tone across the face, fine facial hair, the off-white of a real eye and a catchlight shaped like the actual light. The result is smooth, even and waxy. Un-Plastic puts those details back without redrawing the person.

Will it change the person’s face?

Not in testing. Both blocks tell the model to keep the identity, expression, gaze, pose, crop and light direction. Block A held the face, the smile and the gaze on all ten photos in Nano Banana Pro at 4K, and Block B held the face on all ten in GPT Image at 1K. At 2K, check the crop: one GPT Image run came back slightly tighter.

Why add an upscaler?

The edit block adds what the upscale cannot: pores, sheen and real eyes. The upscaler adds what the block cannot: overall photographic crispness. In testing, Magnific Precision V2 on its own made photos sharper but never added a pore, and the block on its own stayed a little soft. Block first, then the upscale, was best on all three photos tested this way.

Will I need to pay for anything else?

Only your usual tools. The edit runs on whichever site you already use for Nano Banana Pro or GPT Image. The optional upscale uses Magnific Precision V2, which costs about $0.12 an image on FLORA (October 2026), paid to FLORA from prepaid credits. Un-Plastic does not include tool credits.

Does it work in ChatGPT?

Yes. Block B is written for GPT Image, the model behind image editing in ChatGPT. Attach your photo, paste Block B and send. Its change to the skin is gentler than Block A’s in Nano Banana Pro, and it shows best at larger output sizes.

Does it work on darker skin tones?

Yes. On deeper skin the blocks render glow as specular sheen along the cheekbones, brow and nose, not as brightening, and never grey or ashy. Two of the ten test portraits have deep skin, and both passed.

What won’t it fix?

Illustration, anime and 3D-render styles, faces too small in the frame, group shots, and photos with a heavy beauty filter already baked in. Crop in first if the face is small, and do one face at a time.

Can I get a refund?

Yes, within 30 days, if Block A changed the face, the smile or the gaze on your photo. Reply to your Gumroad receipt with the photo you started from and the result. Before you buy, check “What won’t it fix?” above.

What do I get, and how?

A 12-page PDF guide with both blocks, the upscale settings, the two helper lines and the tested before-and-after pairs, plus a plain text file for clean copy-paste. Pay once through Gumroad and the download is yours straight away, with a copy by email. Updates are free.