Speechmatics
Speechmatics is a speech recognition company headquartered in Cambridge, United Kingdom, founded in 2006 by Tony Robinson.
Profile
Speechmatics converts spoken language into text in real time for voice agents, live captions, and call analytics, supporting multiple languages and deployment environments.
Speechmatics is a speech recognition company headquartered in Cambridge, United Kingdom, founded in 2006 by Tony Robinson. It provides speech-to-text, text-to-speech, and voice agent infrastructure, processing spoken language into text for applications such as live captioning, contact center analytics, and medical documentation. The company raised $62 million in Series B funding in June 2022, led by Susquehanna Growth Equity, with participation from AlbionVC and IQ Capital, bringing total funding to $90.6 million.
In 2025, Speechmatics reported 4x year-over-year growth in real-time usage, 9x growth in voice agent usage, and 15x growth in healthcare AI usage. Customers include AI-Media, which processed over 79 million caption minutes in FY25, and LiveKit, which integrated Speechmatics for over 100,000 developers. The company serves regulated sectors such as financial services, healthcare, and broadcasting, and has developed domain-specific language packs for finance and medicine.
Speechmatics operates across cloud, on-premises, and local hardware deployments, and has partnered with Adobe to deliver on-device transcription for Premiere Pro. The company’s stated mission is to understand every voice regardless of accent, dialect, or demographic, and it has focused on reducing AI bias through training on unlabelled voice data. As of June 2026, Speechmatics is a Series B minicorn with 224 active competitors, including Verbit, Otter.ai, and Deepgram.
Who buys this
- Broadcasters and media companies needing live captioning
- Healthcare providers using ambient scribe and clinical documentation
- Financial services firms for call recording and compliance
- Contact centers deploying voice agents for customer service
- Developers building voice-first applications via API
Publicly disclosed clients
- AI-Media
- Adobe
- LiveKit
- NCI
- Media Track
- Sully.ai
Strengths and what to watch
Strengths
- Low-latency real-time STT: partial transcripts in under 250ms and end-of-speech detection in 400ms, enabling natural conversation flow for voice agents.
- Domain-specific language packs: finance and medical models reduce errors on jargon and technical terminology, with up to 70% fewer errors reported in specialized use cases.
- Multi-deployment flexibility: cloud, on-premises, and on-device options, including a partnership with Adobe for local transcription in Premiere Pro.
Watch for
- Competitive pressure from well-funded rivals: Deepgram, Otter.ai, and Verbit have raised significant capital and target similar enterprise segments.
- Dependence on a few large customers: AI-Media alone accounted for 79 million caption minutes in FY25, indicating potential revenue concentration risk.
- Unproven profitability: as a Series B company with $90.6M total funding, Speechmatics has not disclosed net income or operating margins, and the speech recognition market remains capital-intensive.
Recent moves
Key Information
- Industry
- Speech
- Founded
- 1980
- Employees
- 120
- Headquarters
- Cambridge, United Kingdom
Frequently Asked Questions
What does Speechmatics do?
Speechmatics converts spoken language into text in real time for voice agents, live captions, and call analytics. It supports multiple languages and deployment environments, including cloud, on-premises, and local hardware, serving sectors like broadcasting, healthcare, and financial services.
How accurate is Speechmatics speech-to-text?
Speechmatics achieves low-latency real-time transcription with partial transcripts in under 250ms and end-of-speech detection in 400ms. Its domain-specific language packs for finance and medicine reduce errors on jargon by up to 70%, improving accuracy in specialized use cases.
What industries use Speechmatics?
Broadcasters and media companies use it for live captioning, healthcare providers for clinical documentation, financial services for call recording and compliance, and contact centers for voice agents. Developers also build voice-first applications via its API.
How much funding has Speechmatics raised?
Speechmatics raised $62 million in Series B funding in June 2022, led by Susquehanna Growth Equity, bringing total funding to $90.6 million. The company is a Series B minicorn as of June 2026, with 224 active competitors including Deepgram and Otter.ai.
Can Speechmatics run on-premises or on-device?
Yes, Speechmatics offers cloud, on-premises, and on-device deployment options. It partnered with Adobe to deliver on-device transcription for Premiere Pro, allowing users to process audio locally without sending data to the cloud.
Who are Speechmatics' main competitors?
Speechmatics faces competition from well-funded rivals like Deepgram, Otter.ai, and Verbit, which target similar enterprise segments. As of June 2026, it has 224 active competitors in the speech recognition market, which remains capital-intensive.
Sources
- www.speechmatics.com — Company overview, product offerings, customer logos, and case studies.
- www.speechmatics.com — 2025 revenue metrics: 4x real-time usage growth, 9x voice agent growth, 15x healthcare usage growth, 30 million minutes returned to clinicians.
- www.speechmatics.com — $62 million Series B funding in June 2022 led by Susquehanna Growth Equity, total funding $90.6M.
- tracxn.com — Founded 2006 by Tony Robinson, Series B stage, 224 competitors, valuation and investor details.
- www.speechmatics.com — 2026 outlook, deepfake fraud surge, healthcare ROI metrics, and domain-specific accuracy improvements.
- www.speechmatics.com — Finance-specific language pack and entity formatting for financial services.