I was working on a Japanese work email last year, translating sentence by sentence, when I realized I had no idea whether my phrasing sounded junior or senior, distant or warm. Google Translate gave me one answer with no signal about where it sat on any of those axes. I sent the email and never found out if it landed wrong.
That's the gap konid fills. For anything you want to say, it returns three options — casual, neutral, formal — with the register explained and a cultural note on what each choice signals to a native speaker. Not a list of synonyms. An explanation of the relationship the phrase implies.
Here's what that looks like for a concrete case. Say you want to tell someone in Japanese "I missed you today." A literal tool gives you one sentence. konid gives you:
- Option 1 (casual): what you'd text someone you're close to
- Option 2 (warm/neutral): usable across most relationships without risking sounding cold
- Option 3 (formal/tender): appropriate when the emotional weight matters and you want to be precise
Each option includes a note on when it fits and what it signals. Audio pronunciation plays directly through your speakers — no API key required, handled via node-edge-tts.
The install in Claude Code is one line:
claude mcp add konid-ai -- npx -y konid-ai
It also works in Cursor, VS Code Copilot, Windsurf, Zed, JetBrains, and Claude Cowork. If you're on ChatGPT with Developer mode, the endpoint is https://konid.fly.dev/mcp.
Supports 13+ languages: Mandarin, Japanese, Korean, Spanish, French, German, Portuguese, Italian, Russian, Arabic, Hindi, and more.
I built this because I was translating work emails in a language I'm still learning, and every translation that just got me through the task taught me nothing. Seeing three options side by side — with the cultural delta explained — is what actually built my intuition over time.
MIT licensed. https://github.com/robertnowell/konid-language-learning
Top comments (0)