Realtime SaaS

<h2>New</h2><ul><li><p>New field <code>orchestrator_version</code> in the <code>RecognitionStarted</code> message which uniquely identifies the underlying engine version</p></li></ul><h2 style="text-align: start"><strong>Improvements</strong></h2><ul><li><p>Transcription latency reduced by 150-200 milliseconds for English (Enhanced and Standard operating points)</p></li></ul><h2 style="text-align: start"><strong>Fixes</strong></h2><ul><li><p>"Kyiv" output consistently in English transcription</p></li><li><p>Fix for delay in emitting Final transcripts for a subset of languages when using config containing <code>punctuation_overrides={"permitted_marks":[],...</code></p></li><li><p>When the server receives an <code>EndOfStream</code> message, all <code>AddAudio</code> messages received are dropped, and on the first <code>AddAudio</code> received after <code>EndOfStream</code>, the server will send an error message to the client (but not close the connection)</p></li></ul><p></p>

2025.02.13 - Real-Time SaaS

Help Center

Batch SaaS

Batch Container

Realtime Container

Batch Appliance

Realtime Appliance

GPU Transcription Inference Container

GPU Translation Inference Container

Language ID Container

Realtime Kubernetes

speechmatics

In Review

Planned

In Progress

Completed

Rejected

High Priority

Low Priority

Main Roadmap

Quarterly Roadmap

Monthly Roadmap

Tell speechmatics how they could make the product more useful to you.

Have something to say?

Hey {name|there}! 👋

Speechmatics

2025.02.13 - Real-Time SaaS

New

Improvements

Fixes