2021 Review – Accomplishments in Voice-to-Text

As the publication of our Voice Report 2022 looks to the future, we’d love to share some of the major accomplishments of Speechmatics in 2021.

It’s been an incredible year for all areas of Speechmatics. We’ve expanded our team and delivered fantastic improvements in our core speech recognition engine. This has enabled our accuracy levels to reach new highs. Here are just a few of our highlights.

• In October, we launched our Autonomous Speech Recognition. The most powerful, accurate, and inclusive speech recognition ever released, available across every language, for every voice.

• This launch garnered global press coverage, including reports by BBC, TechCrunch, and CNBC, to name a few.

• We were Gold sponsors for three AI Summits (London, New York, Santa Clara), where we gave a featured presentation on AI Bias.

• Our social media followers grew by over 65%, year-on-year (all channels combined).

• A.I. Magazine named Speechmatics the number one speech recognition company using ML and AI.

• We launched our impressive free trial, so businesses and customers can see the benefits of Speechmatics before they buy.

• We continued to grow our team – and have more fantastic job opportunities on our website every day.

• Speaking of, our excellent new website launched in December, complete with brand overhaul and new content layout.

Accuracy Improvements

Thanks to our vision for Autonomous Speech Recognition, we saw our accuracy levels soar in 2021. For the first time, we introduced self-supervised learning into our training and the results have been extraordinary.

Product Updates

Our headline release was, of course, our Autonomous Speech Recognition, but we didn’t stop there. We also made a number of improvements across old favorites, with some great new product updates and features introduced too.

Usage API

We’ve given users access to their transcription usage on our SaaS via API calls.

Language Improvements

As part of our improvements for all languages, our Chinese Mandarin now supports Traditional and Simplified Chinese.

Our API has been updated to allow customers to request a standard or enhanced language model for all our supported languages across all our product deployments

Cancel/Delete Jobs

Rather than let jobs finish or time out, users can now stop and delete running transcription jobs on our batch appliance and SaaS deployments. A huge time saver for any jobs sent in error.

Improved Speaker Diarization

Included in our SaaS, we continue to improve the ability to differentiate between two or more speaking voices.

Alignment

Our alignment product is now available for our SaaS deployment. Users can submit audio files and transcripts, and receive an aligned transcript with timestamps for each word.

Where next? 

With a number of excellent product updates in the pipeline, we’re just as excited about what the new year will bring and what we can achieve in 2022.