Yandex Brand Voice Lite

The main articles are:

• Speech Recognition (Technology, Market)
• Speech technology: On the path from recognition to understanding
• Speech synthesis

2025: Brand Voice Lite Performance

Yandex B2B Tech has launched Brand Voice Lite, a rapid speech synthesis service. The company announced this on July 11, 2025. Now companies just need to upload 20-40 minutes of announcer speech recording to the service to create a unique voice for their brand. It can be used to create personalized voice bots in contact centers, as well as voice acting podcasts, educational materials and other creative business projects. The service is already available to companies on the Yandex Cloud platform.

To synthesize a voice, it is enough to voice the text in the service interface or upload a ready-made entry to it. Announcers can be, for example, guest actors or business owners and ambassadors. They can synthesize voice in different roles: for each additional emotion (for example, sad or aggressive speech), you need to record from 20 minutes. The synthesis model will become available within seven days after uploading records to the service.

Brand Voice Lite technology complies with Yandex principles when working with speech synthesis. A business must obtain a person's consent to use a recording of their voice. Only the customer will have access to the synthesis results. At the same time, he agrees with the rules for using the service, which limit the scenarios for the use of synthesis.

According to Fortune Business Insights, the global volume of the conversational market AI in 2024 reached $12.2 billion, and in 2025 it will grow by 21% to $14.8 billion. The volume of the Russian dialogue market in 2024 was supposed to reach 7.45 billion rubles, predicted in. Naumen

Synthesized voices are most in demand among banks, large retailers and telecom operators, which use them to automate the first support line. In total, with the help of SpeechKit Brand Voice technologies, companies have already synthesized more than 50 voices. After the launch of the Lite version of the service, we expect a multiple increase in such projects, "Ivan Puzyrevsky told CTO of the Yandex Cloud platform.

Brand Voice Lite is based on Yandex SpeechKit's own ML models, which are completed for each client on their datacets. Training data is stored in the company's data centers, which meet the most stringent standards in the field of information and physical security.