🌐 Other languages

🌿 Mint

Minimalist AI Translation CLI — Simple. Fast. Intuitive.

Mint is a single-binary, LLM-powered translation CLI. Set two environment variables and translate anything from the command line — files, piped output, or inline text. Built-in language detection, grammar correction, streaming output, and multi-language rotation.

export MINT_PROVIDER=google-genai
export MINT_API_KEY=your_key

mint -t ja "Good morning"         # おはようございます
echo "早安" | mint -t en          # Good morning
cat document.txt | mint -t fr     # translate a whole file

✨ Why Mint?

• Zero-config — Single binary; API keys via env vars, no config file pollution
• Multi-provider — Google Gemini, OpenAI, Anthropic, or any OpenAI-compatible endpoint (Ollama, LM Studio, OpenRouter, Groq, DeepSeek, llama.cpp, …)
• Smart detection — Auto-detects language on every call; language-neutral content (numbers, symbols) passes through unchanged
• Smart correction — Same-language input? Auto-corrects grammar & spelling instead of translating
• Streaming — Output streams in real-time, no waiting for long translations
• Composable — Pipe-friendly stdin/stdout; pairs seamlessly with grep, sed, xargs, and friends
• Secure — Untrusted input is isolated from model instructions via system/user message separation and per-request random-nonce delimiters; translating adversarial content cannot hijack the LLM's behavior

📋 Installation

Automated install (recommended)

macOS / Linux

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

Auto-detects OS and architecture (Linux/macOS, x86_64/arm64), installs to ~/.local/bin. Override with MINT_INSTALL_DIR or pin a version with MINT_VERSION=v1.0.0.

Windows (PowerShell)

irm https://raw.githubusercontent.com/min0625/mint/main/script/install.ps1 | iex

Auto-detects architecture (x86_64/arm64) and installs to $HOME\.local\bin. Override with $env:MINT_INSTALL_DIR or pin a version with $env:MINT_VERSION = 'v1.0.0'.

Homebrew (macOS / Linux)

brew install min0625/tap/mint-ai

pipx

pipx install mint-ai

npm

npm install -g mint-ai

Manual download

Download the pre-built binary for your platform from GitHub Releases, move it into a directory on your PATH, then verify:

mint --version

🚀 Quick Start

1. Set your provider

# Google Gemini (free tier available — https://aistudio.google.com/apikey)
export MINT_PROVIDER=google-genai
export MINT_API_KEY=your_gemini_api_key

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

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

# Ollama (no API key needed)
export MINT_PROVIDER=openai
export MINT_BASE_URL=http://localhost:11434
export MINT_MODEL_NAME=qwen2.5:7b  # use any model loaded in Ollama

# LM Studio (no API key needed)
export MINT_PROVIDER=openai
export MINT_BASE_URL=http://localhost:1234
export MINT_MODEL_NAME=lmstudio-community/Qwen2.5-7B-Instruct-GGUF  # use any model loaded in LM Studio

# llama.cpp llama-server (no API key needed)
export MINT_PROVIDER=openai
export MINT_BASE_URL=http://localhost:8080
export MINT_MODEL_NAME=qwen2.5:7b  # match whatever model llama-server has loaded

# OpenRouter (one key, hundreds of models — https://openrouter.ai/models)
export MINT_PROVIDER=openai
export MINT_BASE_URL=https://openrouter.ai/api
export MINT_API_KEY=sk-or-...
export MINT_MODEL_NAME=openai/gpt-4o-mini

# Groq (fast inference, free tier)
export MINT_PROVIDER=openai
export MINT_BASE_URL=https://api.groq.com/openai
export MINT_API_KEY=gsk_...
export MINT_MODEL_NAME=llama-3.1-8b-instant

# DeepSeek
export MINT_PROVIDER=openai
export MINT_BASE_URL=https://api.deepseek.com
export MINT_API_KEY=sk-...
export MINT_MODEL_NAME=deepseek-chat

2. Translate

mint --target ja "Good morning"
mint -t zh-TW "Good morning"

echo "The quick brown fox" | mint -t fr
cat document.txt | mint -t zh-TW

Use --verbose / -v (or MINT_VERBOSE=true) to print diagnostic info and token usage to stderr:

mint -t ja -v "Good morning"
# [mint] provider: google-genai
# [mint] single target — skipping language detection
# [mint] target language: ja
# おはようございます
# [mint] tokens: 113 in / 2 out

Typical token usage (measured on gemini-3.1-flash-lite):

Token counts scale with input length. Output tokens vary by target language — Japanese and Chinese tend to produce more tokens than English for equivalent content.

How far does 1M tokens go? (input + output combined, derived from the measured usage above):

Counts combine input and output tokens. Providers price input and output separately and many offer free tiers — check your provider's pricing page for current rates. Google Gemini's free tier at Google AI Studio needs no credit card.

Force the source language with --source / -s to translate input that is also valid in the target language (cross-language homographs, romanized text):

mint -s fr -t en "pain"          # French → bread (without -s, treated as English "pain")
mint -s ja -t en "konnichiwa"    # romaji Japanese → hello

3. Smart language detection

Translation with auto-detection:

export MINT_TARGET_LANG=en

mint "早安"   # Detects Chinese → Good morning

Grammar & spelling correction — when input language matches the target, Mint corrects instead of translates:

export MINT_TARGET_LANG=en

mint "Good mooorning"          # Detects English → Good morning
mint "She don't know nothing"  # Detects English → She doesn't know anything
mint "i luv coding"            # Detects English → I love coding

Language rotation — translates to the next language in the list, wrapping around:

# Two languages
export MINT_TARGET_LANG=en,zh-TW
mint "Hello"   # en → zh-TW: 你好
mint "你好"    # zh-TW → en: Hello

# Three languages
export MINT_TARGET_LANG=en,zh-TW,ja
mint "Hello"       # en → zh-TW: 你好
mint "你好"        # zh-TW → ja: こんにちは
mint "こんにちは"   # ja → en: Hello

🔑 Environment Variables

🚩 CLI Flags

📅 Roadmap

• Multi-LLM provider support (Google Gemini, OpenAI, Anthropic, or any OpenAI-compatible endpoint)
• Smart language detection and multi-language rotation via MINT_TARGET_LANG
• Explicit target language via --target / -t flag
• Explicit source language via --source / -s flag
• Streaming output
• GoReleaser multi-platform binary release (Linux / macOS / Windows)
• Batch translation mode — long input is split at paragraph boundaries and translated chunk by chunk
• Glossary / custom dictionary support
• Output format options (plain text, JSON, Markdown)
• Caching for repeated translations

📄 License

Apache License 2.0 — see LICENSE for details.