German TTS audio on tts.audio runs on Microsoft MAI Voice 2, with two neural voices: Klaus (male) and Mia (female). Both produce Standard German, the High German variety used in news broadcasts, audiobooks, and corporate training across Germany, Austria, and German-speaking Switzerland. Klaus carries all eighteen MAI expressive styles, making him a strong choice for long-form narration where you want subtle emotional range. Mia fits conversational use, ads, and character dialogue. Use cases we see most often: German-language podcast drafts, e-learning localization for the DACH market, accessibility audio for German-language documentation, and pronunciation reference for language learners. For Austrian or Swiss German dialect output, MAI does not have dedicated voices, but you can prompt Gemini with dialect text and it will read it back in a close approximation. PDF upload handles umlauts (ä, ö, ü) and the Eszett (ß) without configuration.