Releases: voicegain/platform
Releases · voicegain/platform
Release 1.2.1_beta
Maintenance release in order to fix issues with Preemptible GPU instances.
Major Features and Improvements
- Flexible mechanism to allocate various types of GPUs to ASR instances. Circumvents the problem with certain GPUs temporarily not being available in the Cloud.
Known issues
- Edge deployment is still in Alpha status. Disabled by default. Can be enabled per Customer Account upon request.
- Cloud IVR has not passed full test suite yet. Disabled by default. Can be enabled per Customer Account upon request.
- Audio Sender Daemon does not support encryption yet.
Breaking Changes
- Broadcast Websocket functionality not longer part of the default set of features.
Bug Fixes and Other Changes
- bug #178: NullPointerException caused by an invalid request for /asr/recognize
- bug #177: Fixed incorrect Uptime computation for Voicegain ASR
- enhancement #176: Added Announcements on home page
Release 1.2.0_beta
This release adds fine-grained control of features available to various types of accounts.
Major Features and Improvements
- More convenient Billing status info and control from within the Portal.
- Fine-grained control of features enabled for each account. This allows Voicegain to control access to features that still are still in limited relase to certain accounts only.
- Enhanced Admin/Account Provisioning Portal - password reset, referrer/referral, premium features, etc.
Known issues
- Edge deployment is still in Alpha status. Disabled by default. Can be enabled per Customer Account upon request.
- Cloud IVR has not passed full test suite yet. Disabled by default. Can be enabled per Customer Account upon request.
- Audio Sender Daemon does not support encryption yet.
Breaking Changes
none
Bug Fixes and Other Changes
- bug #167: Release Notes not being displayed.
- bug #170: Fixed slow redraw of modified words in CC-App
- bug #171: Login failing due to case difference is user email
- enhancement #173: Favicons different between various apps and prod/QA/dev
- enhancement #174: CC-App supports Account landing page
Release 1.1.0_beta.2
This release adds functionality that did not make it into the initial release. It also fixes several bugs.
Major Features and Improvements
- Added ASR engine resource status monitoring to Voicegain Portal UI
- Added Password strength checking
- Added Release Notes info to Voicegain Portal UI
- All three billing styles (manual-refill, auto-refill, invoice) now supported for Enterprise Voicegain customers. Speech.Works customers are created with manual-refill style.
- Provisioning&Admin Portal shows 3 new account fields: referrer, referral, billingStyle.
Known issues
- Edge deployment is still in Alpha status - please contact us before attempting Edge Deployments
- Cloud IVR has not passed full test suite yet. Also, it lacks documentation. Please contact us if you would like to test Cloud IVR.
- Audio Sender Daemon does not support encryption yet.
Breaking Changes
none
Bug Fixes and Other Changes
Most of the bugs fixed in this release are related to login/authentication and to working with Language Models.
- bug #142: Provisioning Portal times out after login.
- bug #143: ASR resets occasionally in the middle of transcription. Was due to empty request to NLM.
- bug #144: Password lockout was mistakenly set to occur only after one mistake. Has been increased to 5.
- bug #145: 401 errors when logging in after a period of inactivity
- bug #147: Service name at the end of the URL in the password (re)set email incorrect.
- bug #148: New Language Models not shown unless user has Manager role
- bug #149: Language Model created as Built-in shows up as User model
- bug #153: Bad redirect from Password Set Dialog
- bug #158: Language Model card for a Context is not showing all the models
- bug #160: Building a Built-In model fails
- bug #161: Ensure correct results are returned from Language Model Query (built-in models, etc.)
- bug #162: Transcription with built-in Language Models fails at initial checks
- bug #163: Offline transcription with built-in Language Model not working
- minor bugs #150, #157
Release 1.1.0_beta.2
Release 1.1.0_beta.2
This release adds functionality that did not make it into the initial release. It also fixes several bugs.
Major Features and Improvements
- Added ASR engine resource status monitoring to Voicegain Portal UI
- Added Password strength checking
- Added Release Notes info to Voicegain Portal UI
- All three billing styles (manual-refill, auto-refill, invoice) now supported for Enterprise Voicegain customers. Speech.Works customers are created with manual-refill style.
- Provisioning&Admin Portal shows 3 new account fields: referrer, referral, billingStyle.
Known issues
- Edge deployment is still in Alpha status - please contact us before attempting Edge Deployments
- Cloud IVR has not passed full test suite yet. Also, it lacks documentation. Please contact us if you would like to test Cloud IVR.
- Audio Sender Daemon does not support encryption yet.
Breaking Changes
none
Bug Fixes and Other Changes
Most of the bugs fixed in this release are related to login/authentication and to working with Language Models.
- bug #142: Provisioning Portal times out after login.
- bug #143: ASR resets occasionally in the middle of transcription. Was due to empty request to NLM.
- bug #144: Password lockout was mistakenly set to occur only after one mistake. Has been increased to 5.
- bug #145: 401 errors when logging in after a period of inactivity
- bug #147: Service name at the end of the URL in the password (re)set email incorrect.
- bug #148: New Language Models not shown unless user has Manager role
- bug #149: Language Model created as Built-in shows up as User model
- bug #153: Bad redirect from Password Set Dialog
- bug #158: Language Model card for a Context is not showing all the models
- bug #160: Building a Built-In model fails
- bug #161: Ensure correct results are returned from Language Model Query (built-in models, etc.)
- bug #162: Transcription with built-in Language Models fails at initial checks
- bug #163: Offline transcription with built-in Language Model not working
- minor bugs #150, #157
Public Beta Release
Release 1.1.0_beta.1
This is a first public release of Voicegain Speech-to-Text Platform.
Major Features and Improvements
This being a first public release, here is a list of key features of the Voicegain platform:
- Cloud based Speech-to-Text API supporting both Transcription (large vocabulary) and Recognition (limited context-free grammars)
- MRCP/VXML/GRXML compatible ASR - tested with Dialogic and Aspect (Voxeo) VXML platforms
- Real-Time transcription - core Real-Time Speech-to-Text engine plus tools like: Audio Sender, Web UI for managing multiple transcription channels, Web based client for displaying live transcriptions
- Voicegain Enterprise Web Portal
- Simplified Transcription only Web Portal (Speech.Works)
Known issues
- Edge deployment is still in Alpha status - please contact us before attempting Edge Deployments
- Cloud IVR has not passed full test suite yet. Also it lacks documentation. Please contact us if you would like to test Cloud IVR.
- Audio Sender Daemon does not support encryption yet.
Breaking Changes
- none
Bug Fixes and Other Changes
- we will begin reporting bug fixes starting from the next release