Yandex SpeechKit hybrid releases

Information about updates to the SpeechKit Hybrid image versions will be published here.

Current versionCurrent version

Release 20.01.23Release 20.01.23

1. Updated all the system components to remove critical vulnerabilities.
2. Built a new envoy container based on Ubuntu 20.04 with the latest security updates.
3. Added the option to enable phone number normalization in speech recognition.
4. Decreased memory consumption in speech synthesis at RPS values exceeding the service's estimated bandwidth

Previous versionsPrevious versions

Release 05.12.22Release 05.12.22

1. Added the option to set logging levels for SpeechKit Hybrid containers.
2. Updated time format in logs to be the same in all components.
3. Fixed some of the issues that prevented agent from launching because of swaydb damage.
4. Added the option to enable TLS termination.
5. Installed the latest security updates.
6. Updated the script to launch services inside containers. Improved system behavior at kill -15: the shutdown is graceful now.