Recognition accuracy in Russian is up to 90%
Client and employee utterances will be placed in chronological order
Stream mode support via gRPC and MRCP protocols
The dictionary can be quickly updated with new vocabulary on any topic and subject area
Recognized text is automatically linked to the speech. Navigate through recordings by clicking on keywords to get to the desired fragment.
Punctuation marks are placed automatically. Numbers are displayed in numeric format, not in script
We can adapt existing language and acoustic models to any subject area to improve recognition quality.
The model is optimized for processing TV broadcasts, news stories, radio shows, podcasts and movies.
The model is adapted to the vocabulary and tasks of the customer and technical support in the field of e-commerce, telecommunications, banking and medical services.
The model is designed to process recordings of telephone conversations on random topics.
The model is also optimized for processing audio recordings made using external microphone, such as interviews.
Wav PCM, 8 kHz /16 bit
Media (broadcast):Wav PCM, 16 kHz /16 bit