Release 1.11.0

This Minor release introduces the RTC Callback API for telephony applications of speech recognition.

Major Features and Improvements

• New RTC API Application mode -- provides tools and APIs to build speech-enabled telephone applications
  • purchase phone numbers
  • attach phone numbers to RTC Apps
  • control flow of you RTC App using Web Callback API
• Guide, Help, and API Spec information has been improved.

Breaking Changes

• Not really a breaking change but names of two Application Modes have changed:
  • Dev API -> STT API (to differentiate from new RTC API)
  • ASR -> MRCP ASR

Known issues

• JJSGF grammar does no support tags yet.
• issue #582: Microphone transcription on Firefox gets stuck (never completes)

Bug Fixes and Other Changes

since Release 1.10.1

• bug #566: Link to Terms of Service not visible in Portal
• enhancement #585: Alphabetical sorting of contexts
• enhancement #583: Direct link to support ticket
• enhancement #567: Separate Guide Card for each App

Release 1.10.0

Major Features and Improvements

• Added basic JJSGF grammar support (no tags yet)
• Changes to Transcription Mode UI
• Microphone capture has better compatibility with different browsers
• Only a single session is billed if running multiple sessions on same audio input

Breaking Changes

• none

Known issues

• JJSGF grammar does no support tags yet.

Release 1.9.0

Major Features and Improvements

• New API to generate short-lived JWT authentication tokens.
• Ability to set allowedOrigins for web API requests to allow cross-origin requests (CORS)
• More responsive TTS - part of IVR functionality coming in 2.0.0
• Real-time Speech-to-Text price is now 1.25 cent/minute

Breaking Changes

• none

Release 1.8.1

This maintenance release provides fixes for the following bugs:

• #534: Transcript Review page stops spinner before the audio file has finished loading
• #530: Last part of incremental word-tree content missing in web api response
• #528: GREG chart tool-tip showing incorrect values

Release 1.8.0

Major Features and Improvements

• New wordTree format for delivering incremental transcription results.
• GREG test grammars can be uploaded in addition to being retrieved from a URL.

Known issues

• Audio Sender Daemon does not support encryption yet.
• Grafana Dashboards configurations are very basic

Breaking Changes

• Preemptible flag is now reserved/internal use.

Bug Fixes and Other Changes

since Release 1.7.1

• bug #518: Details of tiered pricing not being shown
• bug #515: Under-billing for long offline transcriptions
• bug #513: In transcript review - speech audio is colored as if it was mere sound
• enhancement #524: Allow to set the number of rows on one page for GREG experiment view
• enhancement #523: Better GREG Data Upload instruction video

Release 1.7.1

This maintenance release provides fixes for following bugs:

• bug #506: Frequency of timestamps in TXT transcript export too high.
• bug #505: minimimumDelay setting for Live Transcription websocket has no effect.

With this release the free $1.25 credit for new accounts has been reduced to $0.25 but all accounts now get 600 free speech-to-text minutes monthly.

Release 1.7.0

Major Features and Improvements

• Signup has been opened to gmail and other non-business email addresses
• Improved acoustic model for offline transcription.
• Improved efficiency of the core ASR recognizer.
• Many fixes and enhancements around Live Transcription.

Known issues

• Audio Sender Daemon does not support encryption yet.
• Grafana Dashboards configurations are very basic

Breaking Changes

• none

Bug Fixes and Other Changes

since Release 1.6.3

• bug #499: (and related) Session gets stuck in load testing.
• bug #492: Unable to enter person's names with dashes.
• bug #488: OutOfMemory error when processing many large requests with inline audio
• bug #487: Stopped Live Transcription stuck in Online-Starting state
• bug #485: Running Live Transcription with audio.capture flag does not result in working Transcript Preview
• bug #481: Secondary output websocket shows controls enabled even though it is not a control websocket
• bug #479: Unable to change delay for websocket
• bug #474: Unable to change context for websocket
• bug #473: "missing websocket name" error when running REAL-TIME together with SEMI-REAL-TIME sessions
• bug #469: 504 error when submitting file for transcription - timeout
• enhancement #503: IVR-Proxy will now work behind HTTP Proxy.
• enhancement #483: Add a label to each Live Transcription session
• enhancement #482: Improve archives of Live Transcription
• enhancement #478: Improve response of adding a new Websocket

Release 1.6.3

This maintenance release provides an improved acoustic model for off-line transcription.

Release 1.6.2

This maintenance release fixes bug #467

Bug Fixes and Other Changes

• bug# 467: Offline transcription of long files sometimes does not report completion - remaining issues have been fixed

Release 1.6.1

This maintenance release fixes the two Issues reported after Release 1.6.0

Bug Fixes and Other Changes

• bug# 468: IVR-Proxy not starting.
• bug# 467: Offline transcription of long files sometimes does not report completion.