Node Reference
Transcribe Node
Transcribe audio to text using Cloudflare Whisper.
The Transcribe node turns audio into text. Its purpose is speech-to-text: take an audio input — an uploaded file or an upstream audio output — and produce a text transcript that downstream nodes can read.
Not yet available as a canvas node
The Transcribe node is part of the node reference, but it is not currently a node you can add and run on the canvas. There is no Transcribe executor or API route wired into the workflow runtime today, so this page documents the node's intended purpose rather than live behavior, parameters, or limits.
What it is for
Transcribe is the bridge from the audio side of a workflow to the text side. With it you could caption narration, feed spoken input into a text-processing step, or convert generated speech back into text. It takes audio in and emits text out.
- Input: an audio source, such as an Audio node or another node that produces audio.
- Output: the transcribed text.
Provider not finalized in product
The transcription provider for this node is not yet exposed in the canvas runtime. Today, transcription appears only as a declared provider capability rather than a runnable node, so the documentation does not commit to a specific model or provider until the node ships.
Working with audio and text today
Until the Transcribe node ships on the canvas, the related nodes that are available are:
- Audio to upload, play, and pass through audio files.
- Audio Generation to create speech and voice output.
- Text Generation and Prompt for the text side of a workflow.
For the full set of node types and what is available now, see Node types.
Was this page helpful?