Speechmatics

Speechmatics is a technology company specializing in automatic speech recognition (ASR) software. Founded in 2006 as Cantab Research Ltd. by Dr. Tony Robinson, a pioneer in applying recurrent neural networks to speech recognition, the company is headquartered in Cambridge, UK, with additional offices in London, Boston, Chennai, and Brno. Robinson's work in speech recognition dates back to his PhD at Cambridge University in the 1980s, and he has published over 100 research papers on the subject. The current CEO is Katy Wigdahl.

The company develops what it terms 'Autonomous Speech Recognition' technology, utilizing machine learning and deep neural networks to transcribe speech to text. Speechmatics' models are trained on millions of hours of unlabeled audio data, enabling them to understand a wide array of languages, dialects, and accents with high accuracy, even in noisy environments. A key feature is the 'Global English' language pack, a single model designed to comprehend all major English accents. The technology is designed to be inclusive, aiming to understand every voice regardless of demographic, age, gender, or accent.

Speechmatics operates on a Platform-as-a-Service (PaaS) business model, offering its technology to developers and businesses via a flexible API. This allows clients across various industries—including media, contact centers, enterprise communications, and healthcare—to integrate speech-to-text capabilities into their own applications and workflows. Deployment options are flexible, including cloud-based, on-premises, and on-device solutions to meet different security and privacy requirements. Revenue is generated through a usage-based pricing structure, with tiered plans including a free option for smaller workloads and custom enterprise licensing.

The company's product suite includes features such as real-time transcription, speaker identification, automatic punctuation, and translation. In recent years, Speechmatics has released specific products like 'Ursa,' a highly accurate speech-to-text engine, and 'Flow,' an API for building voice interactions. The company has raised a total of $90.6 million over five funding rounds, including a $62 million Series B round in 2022 led by Susquehanna Growth Equity.

Keywords: speech recognition, automatic speech recognition, ASR, speech-to-text, voice technology, neural networks, machine learning, transcription API, real-time transcription, multilingual transcription, speaker diarization, accent recognition, PaaS, deep learning, voice interaction, natural language processing, audio intelligence, Global English, media transcription, contact center analytics