What does Speechmatics do?

Speechmatics provides speech technology and Voice AI for enterprises, offering accurate Speech-to-Text, Text-to-Speech, and Voice Agent solutions. Our models understand every voice and accent across 55+ languages, helping businesses unlock the full potential of voice data.

How accurate is Speechmatics Speech-to-Text?

Speechmatics delivers best-in-market accuracy, achieving up to 99% word accuracy and 96% medical keyword recall in industry benchmarks. Our models handle multiple accents, noisy environments, and multi speakers with ease.

What makes Speechmatics Text-to-Speech different?

Our low-latency Text-to-Speech (TTS) delivers lifelike, human-sounding voices with sub-150ms latency that is ideal for real-time conversations. Developers can stream natural speech in multiple voices and deploy it in the cloud, hybrid, or on-prem for privacy and control.

Can I build real-time voice agents with Speechmatics?

Our voice AI enables developers to build real-time voice agents that listen, understand, and respond naturally. Plug in fast with a flexible API and native integrations to power your AI voice agents.

Which industries use Speechmatics?

Speechmatics is trusted by organizations in media, healthcare, contact center, finance, education, and accessibility. Our technology powers transcription, translation, call analytics, and voice AI applications worldwide.

Speechmatics Extends Transcription Offering With Sounds Feature

Sounds extends the Custom Dictionary feature, which lets users add context-specific words in real-time, such as footballer names or breaking news locations, instantly delivering consistency and reliability of specific words and ultimately increasing the accuracy.

Speechmatics has launched the Sounds feature, a new addition to the company’s current speech-to-text offering. Sounds supports broadcasters by delivering highly accurate transcripts, it allows the speech engine to understand the difference between word pronunciations and the way they are written.

By refining the pronunciation within Custom Dictionary, Sounds can help with the spelling of names, products, acronyms, abbreviations, trademarks, copyrights and alternate word forms.

Ian Firth, VP Products at Speechmatics, explained: “

In the broadcast industry, subtitling for names and words that don’t sound phonetically as they are written is an ongoing bugbear and can be a cause for significant embarrassment by the broadcaster. For example, knowing how to spell Condoleezza Rice’s name flawlessly every time. With Sounds, our engine simply needs pronunciation hints that describe the sounds of the words, like ‘AI’ pronounced as ‘ay eye’. It is not necessary to use phonetic definitions, but something that sounds like the word you want to define and the way that you want it to be written. It can even be used to change things completely, for example you could easily configure it so that ‘Dr’ is written every time ‘Doctor’ is said.”

With other vendors’ offerings, a pronunciation pack is required when using a hints-style feature. With the Speechmatics solution, Sounds is built to enhance the speech engine when required and supporting multiple pronunciations of the same word. As the solution can be delivered in private environments, it is highly secure and accessible only to the broadcaster.

Firth continues,

“Accuracy is still the key metric for speech recognition systems and remains top of mind throughout the development of our solutions at Speechmatics. At Speechmatics, we pride ourselves on unparalleled accuracy rates, and Sounds enables us to continue to improve on our speech-to-text accuracy by building technology that understands pronunciations that are personalised to the user and use case when needed.”

The Sounds feature is available now through Custom Dictionary. To try a demo of the Sounds feature at IBC in Amsterdam 14-18 September, visit Speechmatics at stand 8E17.