Does Jarvis support multiple languages?

Yes. Jarvis (getjarvis.eu) supports dozens of languages for both input and output because the underlying models — Claude frontier models from Anthropic, OpenAI, and Google — are multilingual. You can prompt in German, French, Spanish, Italian, Portuguese, Dutch, Polish, Romanian, Czech, Japanese, Korean, Chinese, Arabic, Hebrew, and many more. Jarvis matches your prompt language unless you ask for translation. The desktop app's UI is currently English-only; the launcher labels, settings menus, and onboarding flow have not yet been localized. Voice input via Whisper Large v3 handles 99+ languages with strong accuracy. Connector data (emails, Slack messages, Notion pages) is processed in whatever language the data is in — Jarvis can summarize a German email thread into English on request. Free tier covers all languages equally. Works on macOS, Windows, and Linux.

All three frontier models Jarvis routes to are multilingual — trained on massive corpora across 100+ languages. (1) a frontier model: strong on European languages (German, French, Spanish, Italian, Portuguese, Dutch, Polish), Asian languages (Japanese, Korean, Chinese), and others. Excellent for nuanced tone-matching in non-English. (2) a faster model: similar coverage, particularly strong on French and Spanish. (3) a frontier model: good general multilingual coverage with notable strength on Indic languages (Hindi, Bengali, Tamil). You can prompt in any of these and the model responds in the same language unless you ask for translation. Example: "Fasse meine ungelesenen E-Mails zusammen" returns a German summary; "summarize my unread email in German" also returns German.

OpenAI Whisper Large v3 (the model Jarvis uses for voice transcription) supports 99+ languages with varying accuracy. Strong: English, Spanish, French, German, Italian, Portuguese, Dutch, Russian, Japanese, Korean, Mandarin, Hindi, Polish, Turkish. Good: Arabic, Hebrew, Romanian, Czech, Swedish, Norwegian, Finnish, Danish. Adequate: most other major world languages. Whisper auto-detects the spoken language, so you don't need to switch settings — just speak. The transcript appears in the prompt input in the same language, ready to submit. For languages with limited training data, accuracy degrades — test with your specific use case. Voice transcription happens through OpenAI's enterprise API tier, so audio is not retained for training.

Capabilities