When one translation isn't enough: building konid for real language use

I wanted to tell my girlfriend 'I missed you today' in Farsi — her first language. Google Translate gave me one answer, no context, no sense of whether it sounded tender or clinical or weird. I had no way to know if it would land the way I meant it.

That's the gap konid fills. For any phrase you want to say, it returns three options ordered casual to formal, with the register explained and cultural context on what separates them. Audio pronunciation plays directly through your speakers — no API key, no copying text into a separate tab — using node-edge-tts under the hood.

The example that made me build this: 'I missed you today' in Farsi has options that range from something you'd text a close friend to something that reads more poetic and deliberate. A single literal translation gives you none of that. Knowing which to pick — and why — is the actual thing you need.

Konid works as an MCP server, so it runs inside Claude Code, Cursor, VS Code Copilot, Windsurf, Zed, JetBrains, and Claude Cowork. Install in Claude Code with:

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

It also works as a ChatGPT app via Developer mode using the endpoint https://konid.fly.dev/mcp.

The tool supports 13+ languages: Mandarin, Japanese, Korean, Spanish, French, German, Portuguese, Italian, Russian, Arabic, Hindi, and more. MIT licensed.

I named it after the Farsi word كنيد — 'do,' as in take action.

https://github.com/robertnowell/konid-language-learning