Transcribe audio

Speechmatics

Speech-to-text API that transcribes recorded files and live audio streams into text with timings and speaker labels.

Rate limits

Batch job creation: 10 new jobs per second recommended; HTTP 429 beyond that9
Batch status requests: 50 per second recommended9
Concurrent realtime sessions: Free 2, Pro 50, Enterprise custom10

Other limits

Concurrent batch jobs: 20,0009
Realtime session length: Ends at 48 hours, after 1 hour with no audio, or after 3 minutes with no audio or ping/pongs10
Custom dictionary size: 20,000 items; more is rejected9

Read from Speechmatics’s own pages on 2026-11-04. Every number links to the page a value came from. For agents, the same facts as markdown: /tools/transcribe-audio/speechmatics.md