Description

Seed Audio 1.5 brings dialogue, music, ambience and sound effects into one audio generation workflow on TopMaker AI. Developed by ByteDance, the model can build a complete sound scene from a prompt, with separate tracks that creators can balance and refine in their editor.

Its controls support detailed performance direction: specify a speaker's emotion and pace, include laughter or sighs, and use timestamp instructions to place voices, effects and music. Text can be combined with audio or video input, and up to six reference recordings can guide the voices.

For projects spanning audio drama, podcast narration, dubbing, localization, and film or game sound design, Seed Audio 1.5 supports generations of up to six minutes across 30 languages and regional variants. The web-based offering uses a freemium pricing model.