Media (images, audio, video)

Transcribe Audio / Video

Speech to text with timestamps, speaker labels, word timings and SRT, VTT, TXT or JSON transcript files, using your own transcription provider.

Transcribe Audio / Video (media_transcribe) turns the speech in an audio or video file into text, timed segments and subtitle files.

The Transcribe step selected in a support call workflow, with Deepgram as the AI provider, the Recording input as the source, Label speakers on, SRT and TXT files and run mode Auto — Transcribe with speaker labels into SRT and TXT files; run mode Auto moves long recordings to the background.

Inputs

AI provider (integration_id): one of the workspace's integrations made from a media-capable connector — see Provider setup. It is chosen in the designer (not an expression). The picker lists only connections that can do the step, and warns when the chosen provider cannot do a feature you picked.

Audio or video (source): A file object (a form's Attachments field such as {{form.recording}}, or a file from another step), a list of files (the first is used), a storage:// reference, an http(s) URL (fetched through the egress guard) or base64 content. Files are read for the run's workspace only — the same rules as the document actions.

Options

Setting	Key	Default	What it does
Language	`language`	detect	A language code such as `en`, `de` or `es-MX`; empty detects it
Label speakers	`diarize`	off	Adds a speaker label to each segment
Max speakers	`max_speakers`	detect	A hint for speaker labelling
Word timings	`word_timestamps`	off	Adds `words` with start and end times
Vocabulary hint	`prompt`	—	Names and terms the provider should expect
Files	`formats`	`["srt"]`	Any of `srt`, `vtt`, `txt`, `json`
Model	`model`	the provider's default	Override the model
Output file name	`output_name`	from the source	Base name of the transcript files
Run	`run_mode`	`auto`	`auto`, `inline` or `background`
Output variable	`output_var`	`transcript`	Where the result goes

Run (run_mode): auto (default) runs the step in the background when the media is longer than MEDIA_INLINE_MAX_SECONDS (300 seconds) or the provider works as an asynchronous job, and the run continues when it finishes; inline always waits in the step; background always hands it to a job. Background needs a saved workflow run — Save & Test always runs inline.

Outputs

Variable	Contents
`{{transcript}}`	`text`, `language`, `duration`, `segments` (`id`, `start`, `end`, `text`, `speaker`, `confidence`, `words`), `words`, `speakers`, `files` (`srt`, `vtt`, `txt`, `json`), `provider`, `model`, `chunks`, `cost`
`{{transcript_text}}`	The full transcript
`{{transcript_segments}}`	Timed segments — loop over them
`{{transcript_files}}`	Every transcript file — attach them to an email
`{{transcript_srt}}`, `{{transcript_vtt}}`, `{{transcript_txt}}`, `{{transcript_json}}`	One file per chosen format

Times are seconds from the start.

Example

Transcribe with speakers

{
  "type": "media_transcribe",
  "label": "Transcribe meeting",
  "config": {
    "integration_id": "<integration-id>",
    "source": "{{form.recording}}",
    "diarize": true,
    "formats": ["srt", "txt"],
    "output_var": "transcript"
  }
}

Then use {{transcript_text}} in a Run Agent prompt to summarise it, or burn {{transcript_srt}} into the video with Media Utilities.

Notes

Long audio is split into chunks when the provider has an upload limit (needs FFmpeg); chunks says how many.
The run history shows the duration, language, segment count and a 200-character preview — never the full transcript.