American English TTS audio on tts.audio uses six MAI Voice 2 voices: Ethan, Grant, Harper, Iris, Jasper, and Olivia. Harper and Olivia carry the full eighteen-style expressive range (cheerful, sad, whispering, angry, and fifteen more), making them the most flexible picks for audiobook narration and character work. Grant, Iris, and Jasper are neutral-only voices, suited to corporate explainers, e-learning, and IVR scripts where you want a steady read. The voices speak General American English, the accent used in most US broadcast media. For British English, Australian English, or other regional variants, switch to the Gemini multilingual model and prompt with the regional spelling and vocabulary you want. PDF upload handles American English documents natively; the text extractor preserves paragraph and sentence structure so the audio paces correctly.