Seed Audio 1.0
The best audio model for anything you make.
AI audio-creation product for generating and editing scene audio, including speech/dialogue, voice-reference based output, sound effects, and music for use in audiovisual workflows.
Recent stories
Creators fed recorded dialogue, Seed Audio, and images into Hailuo H3 for lip sync, UI motion, outfit changes, and reverse-time shots. Hailuo is also pitching H3 for ads, UI/UX, and precise edits.
David Comfort used Seed Audio voices, character sheets, GPT Image 2 setting references, and Seedance 2.0 shots to build lip-sync conversations. He framed it as a workaround for face-reference limits.
A BytePlus thread says Seed Audio 1.0 can create dialogue, effects, background music, and reference voices for clips up to 120 seconds. The workflow can pass audio timing into Seedance for video generation.
David Comfort tested Seedance 2.0 with Seed Audio for three-character lip-synced continuous shots at about $4.50 per 15-second video. Use one blocking event per shot, master-derived close-ups, and lock clauses to improve handoffs.
Higgsfield launched Seed Audio 1.0 with voice replacement, text narration, 18-language dubbing and Claude access through Higgsfield MCP. Early tests and BeatBandit integrations show it being used to audition performances before sending audio-guided shots into Seedance.