November 2026
Speechmatics introduced on-device deployment for its speech-to-text API, enabling ultra-low latency and local data privacy with cloud-parity accuracy. It also l
Speechmatics updated its Realtime API documentation to include support for partial transcriptions, enabling developers to receive real-time, low-latency transcr
Speechmatics updated its custom dictionary feature to include the `sounds_like` option for alternative pronunciations, improving transcription accuracy for ambi
Speechmatics enhanced its diarization feature to separate speakers or channels in audio, now available in Batch, Realtime, and Agent STT (speaker-only). It supp
Speechmatics introduced automatic language identification (Language ID) in its transcription API, allowing jobs to auto-detect the predominant spoken language w
Speechmatics introduced an auto-chapters feature that automatically segments transcribed audio into distinct chapters with titles and summaries via a single API
Speechmatics introduced Agent STT, a turn-based transcription service designed for building voice agents with speaker-labelled, punctuated output. Available onl
October 2026
Speechmatics published a blog post outlining how AI-powered medical transcription reduces clinician documentation burden by converting spoken encounters into te
Speechmatics introduced Oak, a medical speech-to-text model designed for multilingual, multi-speaker clinical consultations. Oak supports 55+ languages, tracks
Speechmatics published a thought leadership piece emphasizing the strain multilingual communication places on global healthcare systems and positioning its Oak
Speechmatics published a guide identifying 10 signs law firms’ transcription workflows are outdated, including slow turnaround, formatting inconsistencies, risi
Speechmatics published a blog post outlining how its AI medical transcription technology supports clinical documentation improvement (CDI) by capturing richer c
Speechmatics published a technical guide emphasizing the need for modular voice agent stacks with specialized components like speech-to-text, orchestration, and
Speechmatics unveiled a 20% accuracy improvement in its real-time speech-to-text engine, now supporting 55+ languages with under 1-second latency. The update ou
Speechmatics published a detailed guide on how to benchmark speech recognition accuracy before committing to an API, emphasizing real-world audio testing over v
September 2026
Speechmatics introduced On-Device speech-to-text, running entirely on end-user machines with no audio sent to servers, achieving within 5% of cloud accuracy and
+110 more updates in the archive, browse by month or year above.
Track Speechmatics on autopilot