Eleven v4 is the fourth generation of speech synthesis from ElevenLabs, the company that sets the standard for voice quality. The model reads text with natural pauses, stress, and breathing, making it hard to tell apart from a human narrator.
Emotions are controlled with tags right inside the text: [whispers] for a whisper, [laughs] for laughter, [excited] for enthusiasm, [pause] for a pause. That lets you create a lively ad, an audiobook with real intonation, or expressive character voiceovers for a game.
Choose from 12 voices — male and female, from soft to announcer-style — and preview each one before generating. The model speaks fluently in English and supports more than 70 languages. A single run converts up to 5000 characters, priced at 10 credits per 1000 characters.