konid

konid (کنید) - Farsi for "do." Take action.

Your language expression wizard. An MCP server for Claude Code that coaches you on natural, socially-aware expression in any language.

This is not a translator. konid is an expression coach: tell it what you want to say, and it shows you how a native speaker would actually say it, with cultural context, tone notes, and audio you can play through your speakers.

Quick Start

Prerequisites

• Claude Code (or any MCP-compatible client, see below)
• Node.js 18+
• ANTHROPIC_API_KEY environment variable set (get one here)

Install

claude mcp add konid-ai -- npx -y konid-ai

That's it. Start a new Claude Code session and ask something like "how do I say 'nice to meet you' in Japanese?"

How It Works

You: "how do I say 'we'll see' in Chinese?"

konid returns 3 options, casual to formal:

1. 再说吧 (zài shuō ba) - casual, slightly evasive, can function as a soft "no"
2. 看情况吧 (kàn qíngkuàng ba) - "depends on the situation," genuinely open-minded
3. 到时候再看吧 (dào shíhou zài kàn ba) - "let's wait and see," most neutral

Plus cultural context: "再说吧 is frequently used as a polite way to decline without directly saying no."

You: "speak 1"

Audio plays through your speakers.

You: "speak 1 slowly"

Slower pronunciation for practice.

Works with

konid runs as a stdio MCP server, which means it works with any client that supports the protocol. The config is nearly identical across tools:

Claude Code

claude mcp add konid-ai -- npx -y konid-ai

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "konid": {
      "command": "npx",
      "args": ["-y", "konid-ai"]
    }
  }
}

VS Code (Copilot)

Add to .vscode/mcp.json:

{
  "servers": {
    "konid": {
      "command": "npx",
      "args": ["-y", "konid-ai"]
    }
  }
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "konid": {
      "command": "npx",
      "args": ["-y", "konid-ai"]
    }
  }
}

Zed

Add to settings.json:

{
  "context_servers": {
    "konid": {
      "command": {
        "path": "npx",
        "args": ["-y", "konid-ai"]
      }
    }
  }
}

JetBrains (2025.1+)

Settings > Tools > AI Assistant > MCP Servers > Add > command: npx, args: -y konid-ai

Tools

coach

Express an idea naturally in a target language. Returns 3 socially-calibrated options with romanization, literal meanings, tone notes, and cultural context.

| Parameter | Description | Default | |-----------|-------------|---------| | text | The idea you want to express (in English) | required | | target_language | Target language | Mandarin Chinese | | context | Social context (e.g. "texting a date", "email to boss") | optional |

speak

Speak text aloud using text-to-speech. Generates audio and plays it through your speakers. Auto-detects the language from the text script.

| Parameter | Description | Default | |-----------|-------------|---------| | text | Text to speak (in the target language) | required | | slow | Slower pace for clearer pronunciation | false | | voice | Edge TTS voice name (override auto-detection) | auto-detected |

replay

Replay the last spoken audio clip.

Supported Languages

| Language | Voice | |----------|-------| | Mandarin Chinese | zh-CN-YunjianNeural | | Taiwanese Mandarin | zh-TW-YunJheNeural | | Japanese | ja-JP-KeitaNeural | | Korean | ko-KR-InJoonNeural | | Spanish | es-ES-AlvaroNeural | | French | fr-FR-HenriNeural | | German | de-DE-ConradNeural | | Portuguese | pt-BR-AntonioNeural | | Italian | it-IT-DiegoNeural | | Russian | ru-RU-DmitryNeural | | Arabic | ar-SA-HamedNeural | | Hindi | hi-IN-MadhurNeural | | English | en-US-GuyNeural |

Any language Claude knows can be coached. The voice auto-detection covers the above, and you can pass a specific voice parameter for others.

Configuration

| Environment Variable | Description | Default | |---------------------|-------------|---------| | ANTHROPIC_API_KEY | Your Anthropic API key (required) | - | | KONID_MODEL | Claude model for coaching | claude-sonnet-4-6 |

Audio Playback

konid plays audio through your system speakers automatically:

• macOS: uses afplay (built-in)
• Linux: tries mpv, paplay, aplay in order
• Windows: uses PowerShell SoundPlayer

If no audio player is available, the file path is included in the response so you can open it manually.

Roadmap

• [x] Remove Python/edge-tts dependency (TTS bundled natively)
• [ ] ChatGPT connector (requires HTTP transport)
• [ ] MCP Registry listing
• [ ] Language-specific coaching plugins (idiom databases, formality registers)
• [ ] Hosted version (no API key needed)

License

MIT