12 May 2026
As seen on Instagram:
www.instagram.com
How to use:
1.
Take a photo with maximum zoom.
2.
Take another photo that is slightly zoomed out (this helps the AI model understand the surrounding area and your subject)
3.
Upload both photos to Gemini and select Pro as your model.
4.
Copy the prompt below, paste it and hit send.
Enjoy!
Prompt:
Task: Reconstruct, upscale, and enhance the provided digitally zoomed-in image using the high-fidelity structural and textural information from the 2 zoomed-out reference images.
Instructions:
Base Composition: Keep the exact framing, perspective, and subject positioning of the zoomed-in image. This is the primary target for the final output.
Detail Extraction: Cross-reference the wider shot to extract true textures, fine details, and accurate colors for the degraded subject matter. Use this data to accurately rebuild the pixelated areas.
Artifact Removal: Completely clean up digital zoom artifacts, interpolation blurring, sensor noise, and chromatic aberration present in the tight shot.
Final Output: Produce a crisp, photorealistic image that looks like it was captured natively with a high-quality optical telephoto lens. Do not hallucinate or invent details that cannot be verified by the reference image.
Instructions:
Base Composition: Keep the exact framing, perspective, and subject positioning of the zoomed-in image. This is the primary target for the final output.
Detail Extraction: Cross-reference the wider shot to extract true textures, fine details, and accurate colors for the degraded subject matter. Use this data to accurately rebuild the pixelated areas.
Artifact Removal: Completely clean up digital zoom artifacts, interpolation blurring, sensor noise, and chromatic aberration present in the tight shot.
Final Output: Produce a crisp, photorealistic image that looks like it was captured natively with a high-quality optical telephoto lens. Do not hallucinate or invent details that cannot be verified by the reference image.