Fix a Blurry Photo

Photo sharpening and restorationPortrait 4:5 · Instagram posts and carouselsE-commerceRestaurants and cafésHotels and spas

Before · input

Input: a white enamel teapot and two tea glasses on a wooden counter, a blurry, low-quality photo — An in-house product photo, deliberately blurred and degraded.AI-generated · made by us

Output: the same teapot and tea glasses in sharp focus, with clear wood grain and beams of light — AI-generated · made by us

Prompt for the model

Restore this blurry, low-quality photo. Make it sharp and clean as if it had been taken in focus with a good camera: recover the edges, remove the blur, JPEG blocks and noise. Keep the main subject - [the white enamel teapot with a black rim and the two tea glasses] - exactly as it is: same shape, same proportions, same colors. Keep the whole scene exactly the same: same objects, same composition, same light and shadows. Do not add, remove or redesign anything; do not invent logos, text or patterns that are not visible. If a detail cannot be recovered, keep it soft rather than making it up. Photorealistic.

Replace the gold [bracketed] parts to match your own product, space or brand.

Keep unchanged

The prompt tells the model not to change these. Check the result anyway: these are the parts that break most often.

Product shape
Colors
Composition
Light and shadow
No invented details that aren't visible

When not to use it

Sharpening doesn't bring back lost detail; the model guesses and redraws it. Don't use it where accuracy matters: text, license plates, documents, faces.

If a texture, pattern or detail the product doesn't actually have shows up in a product page image, don't use that image: customers will find that the product they receive looks different.

What it does and how to adapt it

Sharpens a blurry, low-quality product or venue photo you already have. Upload the photo, paste the prompt and write the main subject of the photo in the square brackets (e.g. "the leather handbag", "the hotel room bed").

The most important sentence of the prompt is the last one: it asks the model to leave unrecoverable detail soft instead of inventing it. The model still redraws some fine details; see the attempt notes below.

The example was generated with Nano Banana 2 via Higgsfield.

Generation log

Model: Nano Banana 2 (Google)
Date generated: October 2, 2026
Attempts: Got it on the first try
Aspect ratio: 4:5
Generated via: Higgsfield API (selected: Nano Banana Pro; job record shows nano_banana_2)
[Fields] in the example: main subject: the white enamel teapot with a black rim and the two tea glasses
Resolution: 2k
Job ID: ab7e99e3-aaec-4288-bad5-2d208e801387

What happened across attempts: Sharpened on the first attempt; shape, color and composition were preserved. The lines on the towel and the wood grain were not legible in the input, so the model redrew them. In other words, the result is not a sharp version of the same photo but a redrawing very close to the original.

Even with the same prompt and the same settings, you won't get an identical result: models generate something different every time. The example shows what kind of result this prompt produces; it is not a guarantee.

If you'll use this image in an adIn Turkey, AI-generated ad content falls under a labeling rule: when a label is required and how to word it.