Bark

Bark is an open-source text-to-audio model from Suno AI (MIT license) that is unique in generating not just speech but the full range of human vocalization — laughter, sighs, crying, gasps, and even background music — all from a single text prompt using audio tags like [laughs], [sighs], [music]. A generative model rather than a traditional TTS system, which means output is creative and variable (not deterministic). Slower than dedicated TTS models but unrivaled for expressive, character-rich audio content. Best for game prototypes, creative audio projects, and applications where emotional expression and non-speech sounds are as important as intelligibility.