5.0.0

Changed

• speechmatics-python has been deprecated in favour of speechmatics-rt for real-time transcription and speechmatics-batch for batch transcription.

• Moved channel field in AddTranscript message to the root of the message.

4.0.9

Changed

• speechmatics-python has been deprecated in favour of speechmatics-rt for real-time transcription and speechmatics-batch for batch transcription.

4.0.8

Changed

• Patch WebsocketClient to work with any websockets version >= 10.0

4.0.7

Changed

• Support requesting a temporary token (JWT) with region and client ref

4.0.6

Fixes

• channel diarization labels defaulting in batch due to a property being in the wrong class.

4.0.5

Fixed

• Missing flag in middleware functions for multichannel mode.

4.0.4

Fix non-multichannel sessions mixing up positional args

Adds support for real-time multichannel transcription using separate audio files.
This feature can be enabled with the --multichannel flag, followed by a comma-separated list of channel names, and an equal number of file paths to transcribe. Each channel name corresponds to the file at the same position in the list.

Example usage:

speechmatics rt transcribe \
  --ssl-mode=none \
  --url=ws://0.0.0.0:9000 \
  --diarization=channel \
  --multichannel=channel_1,channel_2 \
  --lang=en \
  --chunk-size=1024 \
  --raw=pcm_s16le \
  --sample-rate=16000 \
  /path_to_audio/file_1.wav \
  /path_to_audio/file_2.wav

Note: All input files must currently use the same audio format.

4.0.0

Changed

BREAKING CHANGE: Metrics functionality now requires explicit installation

Previously, all metrics dependencies (pyannote, pandas, jiwer, etc.) were
installed by default. This change moves them to an optional '[metrics]' extra
to reduce the default installation footprint.

• Move metrics dependencies to requirements-metrics.txt
• Configure extras_require in setup.py for optional installation
• Add graceful error handling in CLI when dependencies are missing

3.0.6

Support the EndOfUtterance message

3.0.4

Added

New diarization parameter (prefer_current_speaker), as well as exposing speaker_sensitivity for RT mode.