Amazon Transcribe
Fully managed automatic speech recognition (ASR) that converts audio and video into text.
Verified against AWS documentation · 2026-08-24
Key facts
- 01Two modes: batch reads a media file from S3, streaming takes live audio over HTTP/2 or WebSocket
- 02Output is JSON — every word carries a timestamp and a confidence score; SRT/VTT subtitle files are optional extras
- 03Speaker partitioning (diarization) labels speakers sharing one channel; channel identification splits two-channel audio (agent vs caller)
- 04Accuracy tuning ladder: custom vocabulary (free word list) → custom language model (paid, trained on your own text)
- 05Limits: batch media under 4 hours and 2 GB, 100 concurrent jobs (up to 10,000 queued, FIFO), max 2 audio channels
Exam triggers
Phrases and scenarios that point here.
- "Make recordings searchable", "convert calls/meetings to text", "index audio content"
- Subtitles or closed captions for video → Transcribe, not Polly
- Live captioning or real-time agent assist → streaming, not batch
- Strip card numbers or SSNs out of call recordings → Transcribe PII redaction
Not this service
- Text → speech (the reverse direction) → Polly
- Sentiment, entities, or key phrases from the transcript → Comprehend
- Transcript in a different language → Translate; Transcribe only outputs the language that was spoken
Use cases
- 01Turning a backlog of recorded webinars and podcasts into searchable text, so someone can jump straight to the pricing discussion instead of scrubbing a timeline
- 02Contact centre QA — stereo call recordings split by channel so you can measure agent talk time and grep for required compliance phrases
- 03Auto-captioning pipeline for a video platform: upload lands in S3 → Lambda starts a job → SRT file written next to the asset
- 04Redacting PII from recordings before transcripts reach the data lake, so analysts can query them without ever touching card numbers
- 05Live captions for all-hands and training sessions, where remote or hard-of-hearing attendees need text as it's spoken
How it's charged
- 01Billed per second of audio in one-second increments — you pay for media duration, not transcript length, so silence and hold music cost the same as speech
- 0215-second minimum per request — thousands of two-second voice clips each bill as 15 seconds, so concatenate short audio before sending
- 03Tiered monthly volume discounts, with rates varying by Region — choose Region for latency and data residency, then check the rate
- 04Add-ons bill on top of the base rate: PII redaction, custom language models, toxicity detection. Custom vocabulary is free
- 05Two-channel audio bills on total duration, not per channel; Transcribe Medical and Call Analytics sit on separate, higher rate cards
