pvg-rs

Explore, filter, and download artworks from your pixiv bookmarks.

Features

• Sync with your pixiv bookmarks incrementally
• Find or exclude bookmarked artworks by tags, title, or author
• Parallel and reliable download of bookmarked artworks
• Local and remote access via a responsive Web UI
• Cross-platform (Windows, Linux, and Android via Termux are tested)

Highlights

• Custom FTS engine with FM-Index and incremental LSM for instant multi-keyword querying
  • Suffix-based indexing for accurate substring search in CJK titles and tags
  • Optimized with square-root decomposed bitmaps and query planner to achieve <1ms latency on 100K+ collections
• Robust and efficient download manager with asynchronous I/O and file transactions for data integrity
• Minimal resource usage for local-first experience on resource-constrained devices

Screenshot

Getting Started

• Download and extract the latest Release for your platform
• Retrieve your pixiv refresh_token
  • See Retrieving Auth Token or use get-pixivpy-token
• Create a config.json file in the extracted directory
  • A minimal example:

    {
      "username": "<your pixiv username>",
      "refresh_token": "<your refresh token>"
    }

• Run pvg or pvg.exe to start the server
• Open the URL shown in the terminal (Ctrl-click) in your browser
  • Typically http://127.0.0.1:5678

Configuration

The following options can be specified in config.json:

• username: Your pixiv username (required)
• refresh_token: Your pixiv refresh token (required)
• port: Port to listen on (default: 5678)
• host: Interface to listen on (default: "127.0.0.1")
  • Specify a LAN address or "0.0.0.0" to allow access from network
• proxy: Proxy to use for network requests
  • Example: "http://127.0.0.1:8080"
• safe_mode: Turn on to show only illustions with a safe sanity_level (default: false)
• home: Path to the directory to store data and index (default: .)
  • Downloaded artworks are saved in <home>/pix

Build

$ cargo build --release

Credits

The usage of pixiv AAPI refers to the following projects:

• pixivpy
• pixivpy-async