AI Utility

Image Prompt Extractor

Reverse-engineer stunning AI art. Upload an original image to instantly reveal the hidden text prompts, negative prompts, and seeds used to generate it.

Data Copied!

1. Upload Image

🖼️

Drag & Drop Image Here

Supports original PNG, JPG, and WebP files from Midjourney or Stable Diffusion.

2. Extracted Metadata

Waiting for image upload...

AI Image Prompt Extraction Engine

Extract text generation prompts, negative prompts, seed values, CFG scale parameters, and sampler settings embedded in PNG chunks (tEXt/iTXt) from Stable Diffusion, ComfyUI, Automatic1111, and Midjourney outputs. All parsing executes locally in your browser using binary array buffers with zero server uploads.

Frequently Asked Questions

How does this tool find the prompt inside an image?

When AI platforms like Midjourney, Stable Diffusion, or DALL-E generate an image, they embed the exact text prompt, seed number, and rendering parameters into the file's hidden metadata (specifically inside PNG 'tEXt' chunks or JPG EXIF data). Our tool reads the raw binary code of the image to extract and display this hidden text.

Why does it say 'No Metadata Found' for my image?

Many social media platforms (like Instagram, Facebook, and Twitter) and messaging apps (like WhatsApp) automatically strip all metadata from images when you upload them to save space and protect privacy. For this tool to work, you must upload the original, uncompressed image file downloaded directly from the AI generator.

Does this tool upload my private images to a server?

No. HTMLtoPHP utilizes the HTML5 FileReader API. Your image is processed entirely within your device's RAM. We never upload, store, or see the images you inspect.

Which AI generators are supported?

This extractor works flawlessly with native files from Midjourney (v5/v6), Stable Diffusion (Automatic1111/ComfyUI), DALL-E 3, and NovelAI. It reads standard PNG text chunks and standard EXIF parameters.