I translate work emails in Japanese every day. My Japanese is intermediate — enough to read, not enough to know when I'm being accidentally blunt with a client or weirdly stiff with a colleague. Google Translate gives me one output with no signal about whether it's appropriate for the relationship.
The same problem shows up in personal contexts. Register in language isn't just politeness — it encodes the relationship itself. "I missed you today" to a partner you've known for years sounds different from the same phrase to someone you just started dating. The word choice, the verb form, the contraction you use or don't — all of it signals how close you are, how the speaker sees the other person. A single translation strips that out.
So I built konid: for any phrase you want to say, it returns three options ordered casual to formal, with the register explained and a cultural nuance comparison between them. It also plays audio pronunciation through your speakers using node-edge-tts — no external API key needed for that part.
Here's what a query looks like in Claude Code after installing:
claude mcp add konid-ai -- npx -y konid-ai
Then you ask it something like: "How do I say 'I've been thinking about you' in Japanese to someone I'm close with?"
You get back three forms. The casual one uses contracted speech patterns and sentence-final particles that mark familiarity. The middle option is warmer than formal but not intimate. The formal version is what you'd use if you were writing it rather than saying it. Each comes with a note explaining what choosing that option signals — not just grammar, but social meaning.
The same mechanic applies to the work email case. You can ask for "please review this before Friday" in Korean and see whether your phrasing reads as a peer request or a subordinate's ask.
konid supports 13+ languages including Mandarin, Japanese, Korean, Spanish, French, German, Portuguese, Italian, Russian, Arabic, and Hindi. It runs as an MCP server, so it works in Claude Code, Cursor, VS Code Copilot, Windsurf, Zed, JetBrains, and Claude Cowork. It also installs as a ChatGPT app via Developer mode using the endpoint https://konid.fly.dev/mcp.
MIT licensed.
Top comments (0)