Amazon Transcribe for Subtitling

Improve accessibility and usability for media content

Why Amazon Transcribe for Subtitling?

Media content is being generated and consumed at exponential rates. Consumers are enjoying content in a variety of non-traditional environments, for example watching educational videos in the library or enjoying a TV show while commuting home on public transit. Subtitling content helps improve accessibility and engagement by offering customers an option to enjoy content by reading the subtitles. Providing subtitles can also be a complex, compliance requirement for video programming distributors like cable operators and broadcasters to support hard-of-hearing groups. Today, many companies utilize manual transcription services to ensure transcription accuracy of custom terminology and lingo. However, these processes often do not scale and expensive to maintain.

Amazon Transcribe makes it easy for companies to add subtitles to their on-demand and live media content with no prior machine learning experience required. Using a deep learning process called automatic speech recognition (ASR), Amazon Transcribe accurately converts speech into subtitles outputs you need. New customers can get started for 60 minutes free per month for the first 12 months with the AWS Free Tier. 


Make your content accessible to wider audiences including deaf and hard-of-hearing groups, and multi-language support for nonnative users.

Improve engagement by allowing users to enjoy your content, even without sound. Research shows videos with subtitles improve understanding and extend viewing times.

Pay for only what you use. Amazon Transcribe automates your subtitling workloads and scales up and down with your business needs.

Page Topics



Use one service API for both on-demand & live streaming content. On-demand/batch supported formats include FLAC, MP3, MP4, Ogg, WebM, AMR, or WAV. Live streaming supported formats include HTTP2 and WebSocket.

Enable Amazon Transcribe to accurately catch your domain-specific terminology such as names, acronyms, and slang with custom vocabulary and custom language models (CLM). For batch jobs you can use custom vocabulary and CLM to achieve the highest levels of accuracy for Amazon Transcribe.

Improve the live subtitling experience for video broadcasts and in-game chat by controlling the stabilization level of partial transcription results. This gives you the flexibility to display the partial results of a sentence instead of waiting for the entire sentence to be subtitled.

Utilize output batch transcriptions jobs in both WebVTT (*.vtt) and SubRip (*.srt) formats to use as video subtitles for existing workflows. Output files will include any content redaction, vocab filters, and distinguish multiple speakers in both formats.

Filter for specific slang, profanity, or any other inappropriate terminology to maintain quality user experience for all. Create and use multiple vocabulary filter lists to create subtitles appropriate for adult or child audiences based off tags.

Expand the reach of your content by supporting subtitles for multiple languages or translate generated subtitles to localize content with Amazon Transcribe. Amazon Transcribe supports multiple languages, reducing the need for diverse language expertise.