Hey everyone,

I built a stupidly simple model manager with a strict focus on performance, objectivity, and zero bloat.

🎨 GitHub Link: https://github.com/SiliconeShojo/models-info

Here is the full breakdown of everything it brings to the table:

🔥 The Headliners: Pure Quality of Life

• Instant Architecture Filtering: It adds a sleek dropdown menu right to your WebUI header. With one click, you can instantly filter your entire grid by base architecture (SD 1.5, SDXL, Flux, Pony, SD 3.5, etc.).

• Categorized LoRA Folders: It scans your LoRAs and automatically groups them into logical subfolders based on what they actually do: Character, Style, Clothing, Concept, Pose, Background, and Utility.

• Zero-Effort Metadata & Previews: It automatically hooks into Civitai and Arcenciel (with a smart fallback system) to scrape base tags, version labels, and official preview images.

• Disk Space Savior: Instead of bloating your drive with massive uncompressed source images, it automatically downscales and compresses thumbnails into optimized .preview.jpg files. Your UI stays incredibly snappy, and your drive stays happy.

🎛️ Total Control Right on the Cards

You don't just look at the model cards; you actually interact with them:

• One-Click Trigger Words: Click the trigger words displayed right on the card to instantly inject them into your prompt box.

• Direct Web Links: Hit the source badge to jump straight to the model's online page.

• On-the-Fly File Management: You can edit metadata, rename files, or completely delete garbage assets directly from the Extra Networks interface.

⚙️ Deep-Dive Settings & Customization

Under Settings ➔ Models Info, you can fine-tune the backend behavior to match your workflow:

• NSFW Content Filters: Keep your workspace safe (or unsafe) by choosing between four distinct preview filtering levels (None, Soft, Mature, or X).

• API Key Integration: Easily plug in your Civitai API key to fetch data for private, archived, or early-access models.

• Performance Controls: Customize the exact downscaling width for thumbnail images and set maximum concurrent request limits so you don't accidentally get rate-limited by the APIs.

🛑 What it doesn't do (By Design):

To keep the extension lightning-fast and hyper-focused, features like checking for updates, scanning duplicates, or downloading actual model files are not and will not be implemented.

If you need a heavy library manager, this isn't it. If you want speed and objectivity without the extra hassle, give it a spin!

Feedback and contributions are always welcome!