Getting Started with Infinite Image Browsing

This guide will walk you through the installation and initial setup of sd-webui-infinite-image-browsing, a powerful and fast image/video/audio browser for various AI web UIs and a standalone application.

Software Support Overview

Infinite Image Browsing (IIB) offers broad compatibility with popular AI image generation tools:

Installation

You can install and run Infinite Image Browsing in three ways: as an extension for SD-webui, as a standalone Python application, or as a desktop application.

1. As an Extension for Stable Diffusion web UI

This is the most common and recommended way if you use SD-webui.

1. Open the Extensions tab: In your Stable Diffusion web UI, navigate to the Extensions tab.
2. Select "Install from URL": Choose the Install from URL option.
3. Enter the GitHub URL: Paste https://github.com/zanllp/sd-webui-infinite-image-browsing into the "URL for extension's git repository" field.
4. Click "Install": Click the Install button and wait for the process to complete.
5. Apply and Restart UI: After installation, click Apply and restart UI to enable the extension. A new "Infinite image browsing" tab will appear in your SD-webui interface.

2. As a Standalone Python Application

This method allows you to run IIB independently without SD-webui, which is useful for browsing images from other sources like ComfyUI, Fooocus, or NovelAI.

1. Prerequisites: Ensure you have Python installed (3.8+ recommended) and pip available.
2. Clone the Repository: Open your terminal or command prompt and clone the repository:

3. Install Dependencies: Install the required Python packages. The install.py script handles this, including optional dependencies like hnswlib for advanced search features.
   • Note on optional dependencies: If hnswlib or numpy installation fails, you can skip them by setting IIB_SKIP_OPTIONAL_DEPS=1 as an environment variable before running install.py. However, this will disable experimental topic clustering/search features.
4. Run the Application: Execute app.py from the root of the cloned repository.
   By default, it will run on http://127.0.0.1:8000. You can specify a host and port:
   • Integrating with other AI UIs: If you need to view images generated by ComfyUI, Fooocus, or NovelAI, the application's built-in parsers will automatically detect and extract metadata from images in their respective output directories. No special command-line flags are usually needed beyond pointing to the correct image folders (see "Initial Setup" below).

3. As a Desktop Application

A pre-compiled executable version is available for Windows, offering a standalone experience without Python installation.

1. Download from Releases: Go to the releases section↗ on the GitHub repository page.
2. Download the latest executable: Download the appropriate exe file for your Windows system (e.g., InfiniteImageBrowsing-win-x64.zip).
3. Install and Run: Extract the downloaded archive and run the executable.
   • Antivirus warning: Your antivirus software might flag the executable as a false positive. You can generally ignore this. There are usually two compiled versions, with the PyInstaller version having a lower false positive rate.
   • Self-compilation: If you wish to compile it yourself, refer to .github/workflows/tauri_app_build.yml in the repository for instructions.

Initial Setup and Configuration

After installation, a few steps are crucial to get the most out of Infinite Image Browsing.

1. Generating Image Index

For advanced search features to work (e.g., by prompt, model, Lora tags), you need to generate an image index.

• First-time access (UI): When you first open the IIB tab in SD-webui or the standalone application, you'll be prompted to "Generate index for search image" if no images are indexed. Click the button to start. This process can take some time, especially for a large collection of images (e.g., 45 seconds for 20,000 images on an AMD 5600X CPU with PCIe SSD).
• Command-line: For standalone Python applications, you can trigger this on startup:
  (Replace /path/to/sd-webui/config.json with your actual path).

The index is updated incrementally. Subsequent updates check for changes in folders, making the process much faster.

2. Pre-generating Thumbnails and Video Covers

To significantly improve browsing performance, especially for folders with many large images or videos, you can pre-generate thumbnails and video covers.

• Image Cache:
  By default, this only processes --extra_paths (see below). To process SD-webui paths, use --sd_webui_config with --sd_webui_path_relative_to_config.
• Video Covers:

• Verbose Output: Add --gen_cache_verbose to see detailed progress during generation.
• Custom Cache Directory: You can specify an alternative cache directory by setting the IIB_CACHE_DIR environment variable.

3. Adding Custom Search Paths

By default, IIB automatically scans directories configured in SD-webui. You can add more folders for browsing and indexing:

• UI (Recommended): In the application, navigate to the desired folder and use the "+ Add" button (or the right-click menu option) to add it to your "Quick move" paths. You can specify whether to scan it or use it in "walk mode".
• Command-line: For standalone Python applications, use the --extra_paths argument:

4. Desktop Application First-Time Setup

When launching the desktop application for the first time, you may see a modal:

• Read Stable Diffusion Webui Config: If you have SD-webui installed and the IIB extension within it, select this option and point to your SD-webui installation directory. This will allow IIB to directly read SD-webui configurations and share data.
• Skip This Configuration: If you don't use SD-webui or prefer a completely independent setup, choose to skip. You can configure paths and other settings later via the application's global settings.

5. Privacy, Security, and AI Features Configuration (.env file)

For advanced control over security and AI features, you can configure environment variables. Create a .env file in the root directory of the extension/application (you can copy .env.example as a template).

• Secret Key Authentication (IIB_SECRET_KEY):
  • If gradio_auth is enabled in SD-webui, IIB will require a secret key. Define IIB_SECRET_KEY=your_secret_key_here in your .env file. The UI will prompt for this key on access.
• Access Control (IIB_ACCESS_CONTROL, IIB_ACCESS_CONTROL_PERMISSION, IIB_ACCESS_CONTROL_ALLOWED_PATHS):
  • IIB_ACCESS_CONTROL: Set to enable or disable. By default, it's enabled when the service allows public access (e.g., --share, --listen).
  • IIB_ACCESS_CONTROL_PERMISSION: Control permissions (read-write, read-only).
  • IIB_ACCESS_CONTROL_ALLOWED_PATHS: Restrict access to specific paths.
• AI/Natural Language Features (OPENAI_API_KEY, OPENAI_BASE_URL, etc.):
  • To use the experimental Natural Language Categorization & Search features, you must configure an OpenAI-compatible API endpoint and key:
    • OPENAI_BASE_URL: e.g., https://api.openai.com/v1 or your local endpoint.
    • OPENAI_API_KEY: Your API key.
    • EMBEDDING_MODEL: The model to use for embeddings (e.g., text-embedding-3-small).
    • AI_MODEL: The model to use for chat (e.g., gpt-4o-mini).
  • Prompt Normalization:
    • IIB_PROMPT_NORMALIZE=1: Enable prompt normalization to clean boilerplate tags for semantic analysis.
    • IIB_PROMPT_NORMALIZE_MODE: Set to balanced (recommended) or theme_only.
• Cache Directory (IIB_CACHE_DIR): Specify an alternative cache directory for thumbnails and other generated data.

6. Auto Tagging Rules

You can define rules to automatically apply custom tags to images upon indexing:

• UI: Access "Global settings" -> "Auto Tagging". Here you can define rules based on prompt content, model, sampler, etc., to automatically assign custom tags.

Basic Usage Overview

Once installed and configured, you can start browsing your images:

• Infinite Scrolling: Browse large collections of images without paging.
• Search: Utilize various search methods including fuzzy search, advanced tag-based search, and experimental natural language search.
• Walk Mode: Flatten folder structures to browse all images in a hierarchical directory.
• File Operations: Perform basic file operations like delete, move, copy, rename, and create new folders.
• Preview and "Send To": View images in full-screen, access generation information, and send images to other SD-webui tabs (e.g., ControlNet, img2img) or third-party extensions.
• Image Comparison: Compare two images side-by-side with their generation information.
• Batch Download: Package multiple selected images into a ZIP file.
• Multilingual Support: Available in Simplified Chinese, Traditional Chinese, English, and German.

Enjoy a seamless and powerful image browsing experience!