Vatis Tech completes voice recognition technology and reaches 97% accuracy in Romanian language

The speech recognition technology based on artificial intelligence developed by the Romanian startup Vatis Tech (https://vatis.tech/) for the Romanian language far exceeds the accuracy offered by the tech giants’ applications, when transcribing audio-video files.

The platform completed in May, just 9 months after the first investment attracted by the start-up, reaches an accuracy of 97% when transcribing the voice in the text, for the Romanian language. Thus, tens of thousands of files can be processed daily through the Vatis Tech cloud software platform, such as audio-video recordings or real-time broadcasts, radio broadcasts, TV, podcasts, live social media streams. The first two world giants to make such technologies available to the public have an accuracy of less than 90% for the Romanian language.

To date, Vatis Tech has already transcribed 1,147,560 minutes of audio and video recordings and is processing a growing volume each month. Vatis Tech is a company founded by Adrian Ispas, Alexandru Topală and Emanuel Nazare in 2020, three experts in software development. In March this year, Gabriel Mureșanu, an executive with over 25 years of experience, joined the sales development team. The company benefited in September 2021 from a pre-seed financing round, worth 200,000 euros, from several investors (Early Game Ventures, Sparking Capital and Mălin Ștefănescu).

“After the investment, we focused on the development of speech recognition algorithms, especially for the Romanian language, aiming to become the best provider of speech-to-text technology for the local market, both in terms of quality and price. We already have very good feedback from the market. All the companies and institutions that could not benefit from a solution for the very accurate transcription of the voice in the text, now have such a technology at their disposal, being able to integrate it in their own projects. We also aim to actively participate in the digitization and automation process of national companies and institutions. The quality of the Romanian language transcription offered by Vatis Tech certainly brings cost reductions, increased productivity and creates competitive advantages “, says Adrian Ispas, founder and CEO of Vatis Tech.

For the Romanian language transcription, the Vatis Tech start-up managed a long-awaited technological performance in the Romanian market for years. The company is already working with major media providers, both content and monitoring services, and plans to become, in the next period, the technology preferred by most companies and government institutions in Romania. To use Vatis Tech technology, a user only needs to create an online account at https://vatis.tech/ and upload the audio or video files they want transcribed.

The free trial period includes 10 hours of audio or video content that can be converted to text. Subsequently, the technology can be purchased on the basis of subscription packages for a minimum number of hours included, starting from the price of only 1 euro / hour of content or without commitment of hours, for 2 euro / hour, the payment being at the number of hours of transcribed content. One hour of audio or video content is transcribed automatically and quickly in just 15-18 minutes.

The technology is aimed at organizations that have large volumes of audio / video recordings that must be automatically transcribed into text in a very short time for further processing, but also individual users and professionals (teachers, lawyers, journalists, vloggers, presenters). The main beneficiaries of Vatis Tech voice recognition technology are in areas such as media (radio-TV stations, monitoring companies), government, legal (courts, law firms), medical, financial, call center services, but also companies IT (manufacturers or integrators that provide audio-video recording systems or voice-command equipment that require audio-to-text conversion software to perform transcripts of recordings or take instructions and voice commands).

Companies can use the Vatis Tech API for their own transcription and monitoring needs, or they can integrate Vatis Tech technology into their products and services with dedicated support. Vatis Tech is preparing for the second round of financing in the autumn of this year for international expansion. The company plans to develop in the next 18 months new voice transcription algorithms for other 4 – 8 languages ​​in the European area, targeting especially German, French, Spanish, but also in the countries located in Eastern Europe (Czech, Hungarian, Polish ) or in the Nordic area, at the same level of transcription accuracy and affordability.