OpenAI has quietly acquired AI startup Weights.gg, a company known for developing advanced voice replication tools that allowed users to clone celebrity and public figure voices.

According to a report by a famous publication, the acquisition included the startup's intellectual property as well as its small team of employees, though financial details of the deal remain undisclosed.

The acquisition surfaced weeks after Weights.gg officially shut down its platform. Visitors to the company's website are now greeted with a message confirming that all services and content were discontinued on April 1, 2026. The company's farewell statement reflected on the creative community that had used its tools, saying, "From the moment we launched, we watched as creators, artists, and visionaries took our tools and transformed them into something we could never have imagined."

Before shutting down, Weights.gg operated as a social platform where users could create and share AI-generated voice models through its consumer application, Replay. The platform gained attention for hosting realistic voice replicas of globally recognised artists and public figures, including Taylor Swift, Kanye West, members of Blackpink, and political leaders such as Donald Trump and Joe Biden. The tools were also reportedly used to imitate copyrighted fictional characters like Bugs Bunny and Daffy Duck.

Despite the acquisition, OpenAI is not expected to launch a public-facing voice cloning app similar to Replay. Reports suggest that employees from Weights.gg have instead been integrated into multiple teams within OpenAI to support broader AI infrastructure and voice technology initiatives.

The move aligns with OpenAI's growing investment in voice-based AI experiences. The company recently expanded voice capabilities within its developer API, enabling features such as real-time speech translation and conversational AI agents. OpenAI has also integrated ChatGPT into Apple CarPlay, allowing users to interact with the chatbot through voice commands similar to Apple's Siri.

OpenAI has historically taken a cautious approach toward unrestricted voice cloning. Around two years ago, the company revealed it had developed technology capable of generating realistic synthetic voices but chose not to release it publicly because of concerns surrounding misuse and impersonation.