woman — AI-Powered Man Page Translation Tool

English | 中文

woman is a transparent drop-in replacement for the man command. It leverages Large Language Models (LLMs) to automatically translate English man pages into your native language. When run in a non-English locale, it intercepts man queries, translates the raw man page source code via an LLM API, and caches the results for display—all while preserving the original layout, section structure, and groff macros.

Features

• 🌍 Locale-Aware: Detects the system locale and only triggers translation when in non-English environments.
• 🤖 AI-Powered: Uses LLM APIs to deliver high-quality technical translations.
• ⚡ Caching Mechanism: Caches translation results in gzip format to ensure subsequent queries load instantly.
• 🔧 Smart Formatting: Automatically inserts .na (no adjust) directives to prevent groff from adding awkward extra spaces in CJK text.
• 📏 CJK Line Wrapping: Automatically wraps excessively long lines at CJK punctuation marks.
• 🐟 Fish Shell Completions: Includes a command-line completion file for the Fish shell.
• 🔌 Extensible Model Support: Pluggable architecture for model providers (built-in support for DeepSeek, easy to extend).

Installation

Arch Linux

curl -O https://raw.githubusercontent.com/satori-5423/woman/main/PKGBUILD
makepkg -si

Build from Source

Prerequisites

• cargo command (install cargo or rustup)
• man command (man-db or equivalent)
• An API key for a compatible LLM provider

build steps

git clone https://github.com/satori-5423/woman.git
cd woman
cargo build --release
cp target/release/woman ~/.local/bin/woman

Shell Completions (Fish)

cp completions/woman.fish ~/.config/fish/completions/

Configuration

The configuration file for woman is located at ~/.local/share/woman/config.json.

Set Model and API Key

# List available models
woman model list

# Set the active model
woman model set-model deepseek-v4-flash

# Set the API key
woman model set-key your-api-key-here

# (Optional) Set a custom API URL
woman model set-url https://api.deepseek.com/v1

# (Optional) DeepSeek V4 thinking mode
# DeepSeek V4 models reason by default, which makes translation slow and
# expensive for small pages. Thinking is DISABLED by default in woman.
# Re-enable it (and optionally tune the effort) if you want it:
woman model set-thinking enabled        # or: disabled
woman model set-effort low              # low | high | max

View Current Configuration

woman model show

Example output:

Active Model: deepseek-v4-flash
API URL:      https://api.deepseek.com/v1
API Key:      sk-••••••••abcd

Development

Running Tests

cargo test

Adding a New Model Provider

1. Create src/models/<provider>.rs
2. Implement the ModelProvider trait
3. Register the provider in src/models/mod.rs
4. Add the provider name to list_providers()

License

MIT License