Launched this week
Seed Audio 1.0
AI audio from text, voice, images, and references
10 followers
AI audio from text, voice, images, and references
10 followers
Seed Audio 1.0 is a browser-based AI audio studio powered by Doubao-Seed-Audio 1.0. Generate natural speech, multi-character dialogue, voice clones, music, sound effects, and ambience from text prompts, reference audio, or an image. Its multimodal workflow makes it easy for creators, podcasters, filmmakers, game developers, educators, and marketers to build complete audio scenes without stitching together separate tools.
The fact that you can pull in an image as a reference and get matching ambience or sound effects out of it is such a clever touch. Most AI audio tools stop at text prompts, so that multimodal angle really stands out in the workflow.
The dialogue generation felt surprisingly lifelike, especially the multi-character scenes where the voices actually seem to react to each other. Got usable ambience for a short film in just a few minutes without juggling other apps.
Honestly the multi-character dialogue feature kind of blew me away, I threw in a quick script with three voices and it actually sounded like a real scene instead of robots reading lines. Going to mess around with the sound effects next.
the multi-character dialogue feature looks really promising for podcast work. one thing that would make it way more useful for me though is a simple timeline view where i can see all the generated clips laid out and drag them to reorder before exporting, rather than juggling them in separate files
Tried the multi-character dialogue thing and honestly it sounds pretty natural, not robotic at all. Kinda wild that you can throw in an image and get matching ambience back.