Action

Transcribe

Audio → text transcript

Transcribe uses AI to turn an audio file (mp3, wav, m4a, and more) into a clean text transcript the rest of your workflow can read.

How it works

This node listens to a recording and writes down what was said. Feed it a voicemail, an interview, a dictated note, or any audio clip, and it hands back the words as text that later steps can summarize, translate, store, or email.

It is the bridge between spoken audio and everything else the platform does with text. On its own it just produces the transcript; the value comes from what you put after it, a Summarize step to get the gist, a Translate step to cross a language barrier, or a write step to file it away.

You can give it a language hint to sharpen accuracy when you know what was spoken, and a short list of names or jargon so it spells them correctly. Clips are capped at 25 MB, so split or compress a long recording before sending it in.

Common use cases

Real situations where this node earns its place in a flow.

Turn voicemails into text

Transcribe a caller's recording so it can be read, searched, and routed like any other submission.

Summarize a recorded call

Transcribe an interview or meeting, then hand it to a Summarize step for the key points.

File spoken notes

Convert a dictated note into text and write it straight into a Dojobase record.

Transcribe in a workflow

A simple example of where this node sits in a flow.

File upload Upload one or more files Transcribe Audio → text transcript Summarize Text → AI summary

What you set up

  • Point it at an audio file, uploaded through a File upload step or produced earlier in the flow.
  • Optionally set a language hint (like en or es) when you know the language.
  • Optionally list names or terms it should spell correctly.

Common questions

What audio formats work?

Common ones: mp3, wav, m4a, webm, mp4, mpeg, and similar. The clip needs to be 25 MB or smaller, so compress or split a long recording first.

Do I need to set the language?

No. It detects the language on its own. Setting a language hint just makes it a little faster and more accurate when you already know what was spoken.