Optional automatic identification of language during transcription, generally referred to as Language Identification. Supported languages are English and Spanish. Contact your sales representative for information about licensing and activating this capability.

Support for multiple substitution lists per stream

Support for dynamic, per-stream substitution rules using the subst_rules tag

Sentiment processing is now performed after substitution and number identification/translation

Support for sentiment customization by using the senti_list and senti_rules tags

Support for dynamic loading and unloading of language models

Improved warning and error reporting

Licensing flexibility and robustness improvements

Support for the NVIDIA CUDA 9 parallel computing platform