SOTA VOX Kit automatically inserts punctuation marks in transcripts. Sentences and proper names begin with capital letters. Thanks to this, working with the text is comfortable, and the transcript is not inferior in quality to manual formatting.
Each transcript is automatically time-stamped for each word, allowing you to quickly find specific fragments in the original audio or link subtitles by timestamp.
You can add new words to the core dictionary to get the most accurate translations of words and phrases related to a specific subject area, such as product names, technical terminology, or names of individuals.
Stream mode allows you to process records in a mode close to real time. The MRCPv2 protocol is supported.
Ability to flexibly customize a list of words or phrases that will be removed from the transcript, such as obscene language, commercial information, or personal data.
Automatic separation of speakers, for example in mono recordings, where the operator and the client are recorded in one channel. The use of the diarization mechanism significantly improves the quality of recognition and the convenience of further work with the text transcript.
