ConfX Global

Speechmatics

Speech-to-text and voice AI APIs in 55+ languages, deployable in the cloud, on premises or on device.

Conversational AI and AI agentsFeatured

About

Speechmatics is a Cambridge, UK speech recognition company founded by Dr Tony Robinson, who pioneered the use of neural networks for speech recognition at Cambridge University in the 1980s. It provides real-time and batch speech-to-text APIs covering more than 55 languages, with speaker diarization and specialist models for voice agents and medical transcription. Deployment options include cloud, on-premises and on-device, and the company states compliance with HIPAA, GDPR, SOC 2 and ISO 27001. Investors include Susquehanna Growth Equity, AlbionVC and IQ Capital. Published customers and partners include Adobe, AI-Media, LiveKit, Stenograph, Media Track, NCI and Prosodica.

Key facts

Headquarters
Cambridge, United Kingdom
Founded
2006
Chief executive
Katy Wigdahl
Founders
Tony Robinson
Regions served
Global
Markets
Conversational AI and AI agents
Website
speechmatics.com

Products and solutions

  • Real-time speech-to-text API with speaker diarization
  • Batch transcription API
  • Agent STT for AI voice agents, built on the Linden model
  • Medical speech-to-text, including the Oak multilingual model
  • Melia multilingual speech-to-text model
  • On-device, on-premises and cloud deployment

Use cases

  • Live captioning and subtitling for broadcasters and events
  • Speech recognition for AI voice agents
  • Contact centre call transcription and analytics
  • Medical and clinical transcription
  • Media monitoring across languages
  • On-device transcription in video editing and court reporting software

Case studies

AI-MediaScaling real-time captioning with voice AI

120 times more content delivered in 2025 than five years earlier

AI-Media, which provides captioning, transcription and translation to broadcasters, governments and education providers, integrated the Speechmatics engine into its live captioning workflows for accuracy and low latency across languages.

Read more ↗
Media TrackCentralising speech-to-text for global media monitoring

20% improvement in transcription accuracy

Media Track monitors more than 3,000 broadcast channels and 2 million print pages a month in over 20 languages. It consolidated several regional speech engines onto Speechmatics.

Read more ↗
AdobeOn-device speech recognition in Adobe Premiere

Near-cloud accuracy with all audio kept on the device

Speechmatics has partnered with Adobe since 2021, when Premiere added speech-to-text. A new on-device model brings near-cloud transcription accuracy to Premiere while keeping audio local.

Read more ↗