Speechmatics

News Summary:

On July 9, 2026, Speechmatics' technology supported Wellcom's solution designed to assist Dutch doctors by reducing administrative time. This AI clinical documentation system records consultations, transcribes them, and generates reports, addressing real-world challenges such as accents, emotional speech, and speech from anxious, medicated, or elderly patients. Previously, on June 23, the company detailed how Adobe Premiere's on-device speech engine was rebuilt using Speechmatics' technology to meet high standards on macOS and Windows, pushing the boundaries of laptop GPU frameworks for audio inference acceleration, with performance compared against Whisper. Earlier in June, on the 17th, Speechmatics launched Melia, a new multilingual speech-to-text model capable of handling code-switching across all 55+ supported languages, making it available in its Portal Batch API and SDKs. This followed a June 16 case study with Humetrix, which emphasized the critical importance of transcription accuracy in clinical intake, where errors in drug names or conditions could lead to serious medical consequences for patients. The company also noted, on June 15, that OpenAI's Whisper, released in 2022, established a new benchmark for on-device speech recognition, significantly influencing developer expectations with its extensive multilingual capabilities and free availability.