Tether AI Research releases open-source end-to-end speech recognition model TranscribePsy

Tether AI Research has released the open-source multilingual automatic speech recognition model TranscribePsy, with approximately 254 million parameters, which supports local operation on devices such as smart glasses and humanoid robots. This model takes approximately 1.65 seconds to transcribe 10 minutes of audio on the iPhone 17 Pro Max, covering 8 to 25 European languages, and is released under the Apache 2.0 license. Tether CEO Paul Adorno expressed his desire for advanced voice AI to run locally on users' own devices, without the need for data to leave the device.