Gragium introduces ai-powered voice design tool for instant synthetic voice creation

Gragium, a Paris-based voice AI company, has launched a new tool called Voice Design that allows users to generate synthetic voices from text descriptions. The tool eliminates the need for reference audio or speaker consent, offering a faster and more flexible alternative to traditional voice cloning. Voice Design is now available via Gradium’s API and Studio platform, with free access on all plans, including the free tier. The system generates 1 to 5 voice candidates in 3 to 5 seconds, based on attributes such as gender, age, accent, and vocal characteristics. A recent blind listening test showed Gradium’s system outperformed competitors like ElevenLabs and Inworld, with a 72.6% win rate across five languages. The tool also includes a technical evaluation benchmark, with an 83.4% prompt adherence score on the InstructTTSEval dataset.