pi-pdf-watermark

Pi Skill — Add customizable text watermarks to PDF files. Center or tiled, with adjustable angle, font size, opacity, and color.

✨ Features

• Two watermark modes: Single centered stamp or full-page tiled repeat
• Fully customizable: Text content, rotation angle, font size, opacity, color
• Auto page-size detection: Handles A4, Letter, and any custom page dimensions
• Cross-platform: Auto-detects Chinese fonts on macOS / Linux / Windows
• Non-destructive: Outputs a new _水印版.pdf file, never overwrites the original

🚀 Install

# Via npm (recommended)
pi install npm:pi-pdf-watermark

# Or via GitHub
pi install git:github.com/ssdiwu/pi-pdf-watermark

📖 Usage

Once installed, just say:

"给这个 PDF 打水印"

The skill will ask you what text to use, then generate the watermarked PDF.

Examples

| What you say | What happens | |---|---| | "给 report.pdf 打水印" | Asks for text → defaults to tiled mode | | "给 report.pdf 打上「机密」水印" | Runs immediately, tiled | | "打一个居中的 30° 水印" | Uses center mode, 30° angle |

Command Line (standalone)

You can also run watermark.py directly without Pi:

# Install dependencies
pip3 install pypdf reportlab

# Basic usage (tiled by default)
python3 skills/pdf-watermark/references/watermark.py input.pdf "模力无限"

# Center mode
python3 skills/pdf-watermark/references/watermark.py input.pdf "CONFIDENTIAL" --mode center --angle 30 --opacity 0.2

# Custom font size & opacity
python3 skills/pdf-watermark/references/watermark.py input.pdf "DRAFT" --mode tile --font-size 36 --opacity 0.1

# Custom output path
python3 skills/pdf-watermark/references/watermark.py input.pdf "内部资料" -o output.pdf

CLI Options

| Option | Default | Description | |--------|---------|-------------| | input | (required) | Input PDF file path | | text | (required) | Watermark text | | --mode | tile | center (single) or tile (full-page) | | --angle | 45 | Rotation angle in degrees | | --font-size | 42 | Font size | | --opacity | 0.15 | Opacity (0–1) | | --output, -o | *_水印版.pdf | Output file path |

🔧 How It Works

User says "打水印"
       ↓
Skill asks: "请问要打什么水印文字？"
       ↓
User replies: "模力无限"
       ↓
Script runs with default tile mode → outputs *_水印版.pdf

Default behavior:

• Mode: Tiled (满铺) — covers entire page at 45°
• Only asks for watermark text, everything else uses sensible defaults
• Switches to center mode only when user explicitly requests it

📦 Dependencies

• pypdf — PDF reading/writing
• reportlab — PDF generation (for watermark layer)

Both are installed automatically when running via Pi, or manually via:

pip3 install pypdf reportlab

🏗️ Package Structure

pi-pdf-watermark/
├── package.json              # Pi manifest + npm metadata
├── README.md                 # This file
├── LICENSE                   # MIT License
└── skills/
    └── pdf-watermark/
        ├── SKILL.md          # Skill definition (Pi reads this)
        └── references/
            └── watermark.py   # Watermark script (CLI-ready)

🤝 Contributing

PRs welcome! Feel free to open issues for feature requests or bugs.

📄 License

MIT © ssdiwu