How to retouch old photos into new and shiny ones!
Discover how to transform your old, faded, or damaged photographs into stunning high-resolution portraits using a single, powerful ChatGPT prompt. Whether it’s a cherished family photo or a rare historical image, this guide shows you how to professionally restore cracks, enhance details, and apply historically accurate colors — all while preserving the subject’s identity and expressions. Learn the exact prompt that makes it possible, and see before-and-after results that feel like stepping back in time. Perfect for digital archivists, history lovers, and content creators looking to breathe new life into vintage memories.
{
"prompt": "Restore and colorize this historical photograph while maintaining perfect facial similarity and original composition. Apply professional photo restoration techniques to repair damages, cracks, and fading. Use historically accurate and natural color tones for skin, clothing, and environment based on the time period. After restoration, add a highly realistic dark wood frame surrounding the photograph. The frame should appear physically real, with deep, rich wood tones, visible natural grain patterns, and subtle variations in color and texture. It should have a handcrafted feel with clean lines and classic proportions, reminiscent of traditional museum-quality portrait frames. Light reflections and shadows on the frame should mimic those found in high-resolution photographs of real wooden objects. The result should closely resemble a high-resolution image of a restored portrait presented within an elegant, handcrafted dark wood frame. The subject's identity, expressions, and key features must remain 100% faithful to the original photograph.",
"parameters": {
"restoration_strength": 0.95,
"colorization_accuracy": 0.92,
"face_similarity_preservation": 0.98,
"detail_enhancement": 0.85,
"noise_reduction": 0.75,
"historical_color_accuracy": 0.85,
"overall_photorealism": 0.99,
"frame_photorealism": 0.98,
"material_texture_detail": 0.97,
"lighting_realism": 0.96,
"sampling_method": "DPM++ 2M Karras",
"sampling_steps": 30,
"cfg_scale": 7.5,
"seed": -1,
"width": 1024,
"height": 1024,
"denoise_strength": 0.75
} 