Home/Integrations/Transcriptions/Amazon Transcribe
Every recording, searchable as text.
AddPipe extracts an audio track from each recording and pushes it to your S3 bucket. A webhook starts an Amazon Transcribe job, and the transcript lands next to the recording it came from.
Update: The AddPipe Platform now has built-in support for audio transcription, which allows you to automatically detect the language of the audio and transcribe it. You can find more information about this feature in the documentation.
name_audio.mp4Audio only, not the whole video
The transcoding engine produces a separate audio track. Transcribing that rather than the mp4 makes jobs faster and cheaper, and nothing is lost.
Search across every recording
A hundred video responses are unusable until they are text. Once transcribed they can be searched, coded, summarised and quoted.
Inside your own AWS account
The audio is already in your bucket and the transcript is written to another. The recording never leaves infrastructure you control.
What transcripts unlock.
Search
Find the three responses out of four hundred that mention a competitor, in seconds rather than an afternoon.
Coding and analysis
Feed transcripts into your analysis tool so open-ended video answers can be themed like text ones.
Accessibility
Captions and a text alternative for every recording you publish, which is a legal requirement in many contexts.
Details.
Which file gets transcribed?
The extracted audio track, name_audio.mp4, produced by the transcoding engine alongside the mp4. It is far smaller than the video, so jobs start faster and cost less.
Do I need to enable anything in AddPipe first?
Two things: the audio track output in the Transcoding Engine, and a push to an S3 bucket. Amazon Transcribe reads its input from S3, so the push is what puts the file where it can be read.
Which event should trigger the job?
video_copied_s3, which fires once the file is actually in your bucket. Triggering on video_converted risks starting a job before the object exists.
How do I get the transcript back?
Amazon Transcribe writes the result as JSON to an output bucket. Handle the completion event, read the transcript text and store it against the recording name.
What languages are supported?
Whatever Amazon Transcribe supports at the time, including automatic language identification. AddPipe places no restriction on this; it delivers the audio.
Can I use a different transcription service?
Yes. The same pattern works for ElevenLabs Scribe and OpenAI Whisper via Replicate. The audio track is a plain file, so any service that accepts one will do.
Related integrations.
Make your recordings searchable.
Webhooks, the audio track output and storage push are all open during the 14-day trial.