Google Cloud Speech-to-Text is a speech recognition API that transcribes audio into text across more than 125 languages and variants. Google brought the API to general availability in 2017, and it now processes more than 1 billion voice minutes each month across customer deployments.
The API accepts both streaming and batch audio, applying automatic punctuation, speaker diarization, and domain-specific models tuned for phone calls or video content, and developers can adapt recognition to custom vocabulary through model adaptation rather than retraining from scratch. A newer V2 API added a Dynamic Batch option aimed at cutting cost for workloads that do not need real-time results.
The service targets developers building transcription, voice-command, or call-analytics features into their own applications rather than end users seeking a standalone app. Google gives 60 minutes of audio processing free every month; beyond that, standard models cost $0.006 per 15 seconds (about $0.024/minute), with enhanced models for phone calls and video priced higher, and new Google Cloud customers also receive $300 in trial credit.
