Speechmatics

What is Speechmatics?

Speechmatics provides speech-to-text APIs for live and recorded audio, aimed at voice AI, captions, meeting notes and contact center analytics. A single model supports 55+ languages, including accents and mid-conversation language switching.

Key capabilities of Speechmatics

• Real-time transcription: Sub-second latency for live audio.
• 55+ languages: One model with accent and language switching support.
• Speaker diarization: Labels who spoke.
• Custom vocabulary: Adds domain terms to improve accuracy.
• Medical Model: A model tuned for healthcare speech.
• Flexible deployment: On-device, on-premises or cloud.

How Speechmatics works

Developers send audio to the API or run Speechmatics on-premises, and receive transcripts with timestamps and speaker labels. Different models trade off accuracy, multilingual coverage, latency and cost. Data is not logged by default.

Who uses Speechmatics?

Developers building voice agents, captioning, meeting tools, legal transcription and contact center analytics use Speechmatics. Healthcare teams can use the Medical Model. The vendor lists ISO 27001, GDPR, HIPAA and SOC 2 Type II, and cites an independent August 2026 test showing a 1.07 percent pooled word error rate.

Speechmatics pricing

Speechmatics lists a Free tier with $100 of credit and no card required, with transcription from $0.12 per hour. Pro uses credit packs from $25 or monthly subscriptions. Enterprise is custom priced for large or regulated deployments.

Speechmatics alternatives

Replicant is a conversational AI platform for contact centers rather than a transcription API. Smith.ai is a virtual receptionist service with human and AI call answering. Listnr generates speech from text, the reverse of transcription.