OpenAI Speech Recognition (SR) Plugin 1.1.0 to the UniMRCP Server (UMS) has been released.
The plugin is based on the following components:
- OpenAI Realtime Transcription WebSocket API
- Libevent 2.1.9
- Rapidjson 1.1.0
The plugin is available for the following UMS versions:
The binaries are currently available for the following Linux distributions:
- Red Hat 8
- unimrcp-openai-sr-1.7.1-1.el8.x86_64.rpm
- unimrcp-openai-sr-1.8.1-1.el8.x86_64.rpm
- Red Hat 9
- unimrcp-openai-sr-1.7.1-1.el9.x86_64.rpm
- unimrcp-openai-sr-1.8.1-1.el9.x86_64.rpm
- Ubuntu 20.04 LTS
- unimrcp-openai-sr_1.7.1-focal_amd64.deb
- unimrcp-openai-sr_1.8.1-focal_amd64.deb
- Ubuntu 22.04 LTS
- unimrcp-openai-sr_1.7.1-jammy_amd64.deb
- unimrcp-openai-sr_1.8.1-jammy_amd64.deb
This release introduces support for various auth methods. The release also provides a few fixes and improvements.
The detailed list of changes introduced in this release follows.
New Features- Added detailed time metrics to RDRs.
- Added
support for various auth methods including bearer token specified in
Authorization header or auth-key specified as header or query param.
Fixed ProblemsConfiguration Parameters- Added a new attribute 'auth-method' to the element 'openai-settings'
Miscellaneous- Updated the Usage Guide to reflect the changes introduced in this release.
Visit the OpenAI SR plugin page for more information.
Thank you for using UniMRCP.
--