Voice cloning startup ElevenLabs today announced a new tool that lets users generate sound effects through prompts. The company first announced the project in February.
The tool, available to all users starting today, allows users to generate sound snippets by inputting prompts such as “sound of waves,” “metal crashing,” “birds chirping” or “race car engine noise.”
The sound effects tool also lets you generate instrumental music clips up to 22 seconds long using prompts like guitar loops, jazz saxophone solos, music techno loops, and more.
Free users can generate 10,000 characters per month. Generating a sound byte takes about 150 characters per request. Generally, free plan users can generate about 60 sound effects per month. Also, when publishing content with sound clips, you must credit the sound to “elevenlabs.io” in the title.
ElevenLabs said it used Shutterstock's audio library, which contains licensed tracks, as a training tool for its models. The company added that users who trialed the tool during its alpha testing phase included video game developers, film producers, social media content creators and marketers.
The startup noted that the tool does not allow sound generation through prompts that violate its prohibited content and usage policy, including topics such as self-harm, threats to child safety, and scams.
While there are only a few companies and startups working on AI-powered sound generation, ElevenLabs may find the music generation field to be crowded: Stability AI-backed Harmonai has released Dance Diffusion; Google is working on MusicLM; OpenAI has Jukebox and Meta has an AudioCraft model; TikTok and Adobe are also experimenting with their own generative AI-based music creation tools.
We're launching our AI newsletter: Sign up here to start receiving it in your inbox starting June 5th.