Infinite Image Browsing Overview

Infinite Image Browsing is a powerful and fast image, video, and audio browser designed for various AI art generation platforms like Stable Diffusion web UI, ComfyUI, Fooocus, NovelAI, and StableSwarmUI. Its primary purpose is to provide an efficient and feature-rich browsing experience for generated media, offering infinite scrolling and advanced search capabilities based on image parameters. It can also operate as a standalone application.

Primary Purpose

The project aims to enhance the media browsing experience for AI art enthusiasts and developers by addressing common limitations in traditional file browsers. It focuses on speed, intelligent organization, and flexible access to large collections of AI-generated content.

Key Features

The application boasts a comprehensive set of features designed for optimal media management:

• Excellent Performance: Achieves millisecond-level display times after initial caching, utilizing thumbnails (default 512px, adjustable) and supporting pre-generation of thumbnails and video covers (--generate_video_cover, --generate_image_cache) for improved speed. Users can control grid image widths from 64px to 1024px and specify a custom cache directory.
• Image Search & Favorite: Converts prompt, model, Lora, and other generation information into searchable tags, sorted by frequency. It includes tag autocomplete, auto-translation, customization, advanced Google-like search, and fuzzy search by filename or generated information. Users can also add custom search paths and favorite images with custom tags.
• View Images/Videos & Send To: Displays comprehensive image generation information, even in full-screen preview. Supports sending images to other tabs and third-party extensions like ControlNet and openOutpaint. Full-screen mode offers custom shortcut keys and navigation. Video files from remote servers are also playable.
• Multiple Usage Methods: Can be installed as an extension for SD-webui, run independently as a Python application, or used as a standalone desktop application. It supports various popular AI software.
• Walk Mode: Automatically loads content from the next folder (similar to os.walk), allowing seamless browsing of thousands of files without manual paging. It flattens folder structures for easier navigation.
• Preview based on File Tree Structure & File operations: Offers file tree-based previews with automatic refreshing. Includes basic file operations like multi-selection for deleting, moving, copying, packing for download, adding/removing tags, and drag-and-drop functionality. A "Keep Multi-Select" button allows persistent selection for multiple operations.
• Image Comparison (similar to Imgsli): Provides a side-by-side comparison of two images, along with their generation information.
• Multilingual Support: Available in Simplified Chinese, Traditional Chinese, English, and German, with a welcoming stance for new language contributions.
• Privacy and Security: Features custom secret key authentication, configurable file system access control (default for public services), customizable allowed paths, and read-only mode options.
• Packaging/Batch Download: Enables downloading multiple images from search results, grid views, or walk mode, with items added via drag-and-drop or "Send To."
• Keyboard Shortcuts: Customizable shortcuts for common actions like deleting and managing tags.
• Natural Language Categorization & Search (Experimental): Groups images by semantic similarity of prompts and allows natural-language retrieval. This experimental feature relies on prompt extraction, normalization, embeddings, clustering, and LLM-based title generation, with robust caching and incremental updates. It uses OpenAI-compatible APIs for AI calls.

Supported AI Software

The project offers broad compatibility across several prominent AI image generation platforms:

• Stable Diffusion web UI: Fully supported (including a Stealth version option).
• ComfyUI: Partially supported.
• Fooocus: Supported.
• NovelAI: Supported.
• StableSwarmUI: Supported.
• Invoke.AI: Supported.
• Pixiv: Supported via the external pixiv_iib_plugin.