Soniox was founded in 2020 with the mission to deeply understand audio and make it universally accessible and useful. In 2021, Soniox gave birth to the world’s first unsupervised learning system for speech recognition, lifting the limitations that used to throttle speech system performance. Shortly after, Soniox launched its Speech AI platform, where each line of code was written from scratch: from proprietary artificial neural networks, to distributed audio processing to real-time inference engine. In June 2022, Soniox released Speaker AI, setting a new bar for recognizing speakers robustly in real-world environments. Today Soniox Speech and Speaker AI are world-leading in accuracy, latency and scalability.
Yet this is only the beginning. To achieve our mission to deeply understand audio, we are now working on language expansion and spoken language understanding, where research is still in its infancy and where Soniox is leading the way to revolutionize conversational intelligence in all facets of our lives.