GenLens shows every Stable Diffusion render in your folder together with its prompt, seed, model and LoRAs. One file. Runs on your machine. Works with ComfyUI, Automatic1111 and Forge.

Which seed was it? Which LoRA weight? Which sampler? The answers are inside the file, but reading them means opening the PNG info tab for every single image. GenLens reads that metadata once, caches it, and puts every answer one click away.
Click any render and read its recipe. Copy the prompt or the seed and go straight back to generating.
Prompt, negative prompt, checkpoint or diffusion model, LoRAs with strengths, seed, steps, CFG, guidance, sampler, scheduler, denoise and latent size.
Save an image’s workflow straight into ComfyUI’s Workflows panel, or queue the exact prompt again with the same seed and settings.
Up to four images next to each other. Differing settings are highlighted and prompts are diffed word by word, so you see what changed between two seeds.
Plain words, quoted phrases and fields like seed:, model:, lora:, steps:>30 or date:week. Filter menus for model, LoRA, sampler, tool, aspect and tags.
F to favorite, D to mark for deletion, 1 to 5 to rate, T to tag. Filter by any of it and move marked images to the Trash in one confirmed step.
Background thumbnails and a metadata cache keep thousands of images quick to open. The grid refreshes itself while you render.
Click any word to see how often you use it and how your best-rated images score with it. Collections, saved views and PNG or PDF contact sheets keep client work organised.
Hide sections you never need, choose dark or light, adjust thumbnail size. Settings are remembered per browser.
A single Python file serving on localhost. No account, no upload, no telemetry. Your renders never leave your disk.
A clean grid, newest first, with a folder tree for your output folders. Favorites, ratings and deletion marks are visible at a glance. Keyboard shortcuts for everything you do often.


Pick up to four images and compare them side by side. Settings that differ light up, and the prompts are diffed word by word. No more guessing which LoRA weight made the difference.

Mark the misses with one key while you browse. Switch to the “To delete” filter, review the list, confirm once. Files go to the Trash or Recycle Bin, not into the void.
GenLens walks the generation graph generically instead of looking for fixed node names, so custom node packs usually just work.
Turn a collection or a filtered view into a contact sheet with file names, seeds and models under each image, as PNG or a multi-page PDF. Or export a zip with the generation metadata stripped.

Copy GenLens into the folder with your images.
The launcher starts a local server and opens your browser.
Click an image, read its recipe, copy what you need.
Python is the only requirement. Everything else is already on your machine.
Dark or light, and only the metadata you care about.


GenLens is in its final stretch before sale. Until then, download the trial: the full program, limited to the 10 newest images of a folder, with export, contact sheets and the ComfyUI actions reserved for the licensed version. When keys go on sale, the same download unlocks in Settings.