mint-ai 0.0.0a12

Minimalist AI translation CLI powered by LLMs

English | 繁體中文

🌿 Mint

Minimalist AI Translation CLI — Simple. Fast. Intuitive.

---

Mint is a lightweight, LLM-powered translation tool for the command line. Choose your favorite LLM provider (Google Gemini, OpenAI, Anthropic, or local Ollama), and get fluent, natural translations instantly with optional smart language detection.

---

✨ Why Mint?

Most translation tools are either too bloated or too locked into a specific platform. Mint is built around a single philosophy: do less, do it well.

• Minimal — One command, no noise
• Fast — Calls the LLM API directly, no intermediate layers
• Multi-provider — Choose between Google Gemini, OpenAI, Anthropic, or local Ollama models
• Flexible — Supports any language pair, freely specify your target language
• Smart detection — Optionally detect input language and auto-translate to your preference
• Composable — Pipe-friendly stdin/stdout design, fits naturally into any workflow

---

📋 Installation

pipx (Python Package Index)

The simplest way to install — Mint is available on PyPI:

pipx install mint-ai

After installation, the command is mint:

mint --version

npm (Node Package Manager)

If you prefer npm:

npm install -g mint-ai

Then use:

mint --version

Automated Install (One-liner)

Downloads the latest binary automatically:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/min0625/mint/main/script/install.sh)"

Features:

• Detects your OS and architecture automatically (Linux/macOS, x86_64/arm64)
• Verifies SHA256 checksums
• Installs to ~/.local/bin by default (override with MINT_INSTALL_DIR)
• Shows PATH setup hints if needed
• Supports pinning a specific version: MINT_VERSION=v1.0.0 bash script/install.sh

go install

If you have Go 1.21+ installed:

go install github.com/min0625/mint/cmd/mint@latest

The binary will be available as mint in your $GOPATH/bin directory (usually ~/go/bin).

Manual Download from GitHub Releases

Download pre-built binaries directly from GitHub Releases:

# Linux x86_64
curl -L https://github.com/min0625/mint/releases/latest/download/mint_Linux_x86_64.tar.gz \
  | tar xz && sudo mv mint /usr/local/bin/

# macOS arm64 (Apple Silicon)
curl -L https://github.com/min0625/mint/releases/latest/download/mint_Darwin_arm64.tar.gz \
  | tar xz && sudo mv mint /usr/local/bin/

# macOS x86_64 (Intel)
curl -L https://github.com/min0625/mint/releases/latest/download/mint_Darwin_x86_64.tar.gz \
  | tar xz && sudo mv mint /usr/local/bin/

# Windows x86_64 (PowerShell)
# Download mint_Windows_x86_64.zip from releases page and extract to a directory in your PATH

Verify Installation

mint --version

---

🚀 Quick Start

1. Choose your provider

# Google Gemini (free tier available)
export MINT_PROVIDER=google-genai
export MINT_API_KEY=your_gemini_api_key
# Get a free API key at: https://aistudio.google.com/apikey

# OpenAI
export MINT_PROVIDER=openai
export MINT_API_KEY=sk-...

# Anthropic
export MINT_PROVIDER=anthropic
export MINT_API_KEY=sk-ant-...

# Local Ollama (no API key needed)
export MINT_PROVIDER=ollama
export MINT_BASE_URL=http://localhost:11434
export MINT_MODEL_NAME=llama2

2. Translate with explicit target language

# Specify a target language (BCP-47 tag) using --target or -t flag
mint --target ja "Good morning"
mint -t zh-TW "Good morning"

# Pipe from stdin
echo "The quick brown fox" | mint -t fr

# Translate a file
cat document.txt | mint -t zh-TW

3. Smart language detection (optional)

Set your target language preference using MINT_TARGET_LANG:

# Single target language
export MINT_TARGET_LANG=en
mint "早安"             # Detects Chinese → translates to en
mint "Good mooorning"  # Detects English → grammar & spelling correction

# Multiple target languages (language rotation)
export MINT_TARGET_LANG=en,zh-TW,ja

mint "Hello"         # English input → translates to zh-TW (next in rotation)
mint "你好"          # Chinese input → translates to ja (next in rotation)
mint "こんにちは"     # Japanese input → translates to en (wraps around)

The tool automatically detects the input language and applies the appropriate transformation.

---

🔑 Environment Variables

Variable	Description	Required	Default
MINT_PROVIDER	LLM provider: google-genai, openai, anthropic, ollama	Yes	—
MINT_API_KEY	API key for the chosen provider	Conditional*	—
MINT_BASE_URL	Custom API endpoint; required for ollama (e.g., for self-hosted or local services)	Conditional*	Provider default
MINT_MODEL_NAME	LLM model name to use	Optional	Provider default**
MINT_TARGET_LANG	Target language(s) - single or comma-separated (e.g. en, en,zh-TW,ja)	Optional	System locale or en

Conditional: MINT_API_KEY required for google-genai, openai, anthropic; not needed for ollama. MINT_BASE_URL required for ollama. Default models: google-genai: gemini-3.1-flash-lite, openai: gpt-4o-mini, anthropic: claude-haiku-4-5; ollama: none (must specify).

Language Resolution Priority

The tool uses the following priority order to determine the target language(s):

1. Flag: --target / -t CLI flag (highest priority)
2. Config: MINT_TARGET_LANG environment variable
3. System: Operating system locale
4. Default: en (lowest priority)

---

🎯 Design Principles

Mint follows the Unix philosophy — do one thing, and do it well.

Principle	Description
Zero-dependency install	Single binary, works out of the box
Multi-provider	Supports major LLM services plus local alternatives
Composability	Pairs seamlessly with grep, sed, xargs, and friends
Transparent output	Results go to stdout, errors go to stderr
Environment-friendly	API keys managed via environment variables, no config file pollution

---

🗺 Roadmap

• Multi-LLM provider support (Google Gemini, OpenAI, Anthropic, Ollama)
• Smart language detection and multi-language rotation via MINT_TARGET_LANG
• Explicit target language via --target / -t flag
• GoReleaser multi-platform binary release (Linux / macOS / Windows)
• Batch translation mode
• Glossary / custom dictionary support
• Output format options (plain text, JSON, Markdown)
• Caching for repeated translations

---

📄 License

Apache License 2.0 — see LICENSE file for details.