Stable Diffusion Prompt Reader

Information beyond the image

AI image-generation tools store metadata in different formats. Once an image leaves its original interface, reading its prompt, seed and generation settings is not always straightforward.

I built a standalone Python tool that separates metadata reading and processing from the image-generation application, with both a desktop interface and a command-line interface.

One reader for multiple formats

The parsing framework supports formats used by A1111, ComfyUI, NovelAI, InvokeAI, Easy Diffusion, Draw Things and other tools. The project documentation tracks the exact scope and limitations.

The desktop application supports drag-and-drop reading, prompt copying, text export, and editing or removing metadata. It runs on macOS, Windows and Linux.

From an application to workflow nodes

The ComfyUI companion reuses the parsing core to bring reading, parameter processing and saving into workflows. Prompt Reader extracts metadata; Prompt Saver and Parameter Generator work together to write additional parameters to output images and improve metadata compatibility across tools and websites.

Sharing the parsing core lets the standalone application and node-based version evolve along the same compatibility roadmap.

Publicly verifiable work

Both repositories are public on GitHub. The star counts shown here are dated snapshots; the repositories remain the source of truth for features, format limitations and releases.