Voice cloning startup ElevenLabs launched a new tool for customers to generate sound results by prompts right now. The corporate first announced the project back in February.
The instrument, out there to all customers beginning right now, permits customers to sort in prompts like “waves crashing,” “steel clanging,” “birds chirping” and “racing automotive engine” to generate snippets of sounds.
The sound results instrument also can generate instrumental musical clips of as much as 22 seconds with prompts reminiscent of guitar loops, jazz saxophone solos and music techno loops.
Free customers get 10,000 character generations per 30 days — a sound byte era takes round 150 characters per request. Basically, free-tier customers can generate almost 60 sound results per 30 days. Additionally they need to attribute the sound to “elevenlabs.io” within the title whereas publishing any content material containing the sound clip.
ElevenLabs mentioned that it used Shutterstock’s audio library containing licensed tracks as a instrument to coach its mannequin. The corporate added that customers that attempted out the instrument through the alpha testing section included online game builders, movie producers, social media content material creators and entrepreneurs.
The startup famous that the instrument doesn’t permit sound era by prompts that violate its Prohibited Content and Uses Policy, which incorporates matters like self-harm, threats to youngster security and fraud.
Whereas there are just a few corporations and startups engaged on AI-powered sound era, ElevenLabs would possibly discover the music era area crowded. Stability AI-backed Harmonai has launched Dance Diffusion; Google has labored on MusicLM; OpenAI has Jukebox; and Meta has its AudioCraft mannequin. TikTok and Adobe have additionally experimented with generative AI-based music creation instruments of their very own.
We’re launching an AI e-newsletter! Join right here to start out receiving it in your inboxes on June 5.