Actions16
- LLM (Text Generation) Actions
- Image Generation Actions
- Video Generation Actions
- Text-to-Speech Actions
- Speech-to-Text Actions
- Music Generation Actions
- Embeddings Actions
- Content Moderation Actions
- Custom Inference Actions
Speech-to-Text → Transcribe
AI-generatedSummary
Transcribe audio to text using a selected Chutes.ai speech-to-text chute or a custom chute URL. Accepts audio input as binary data, a URL, or a base64 string, and returns the transcription result as continuous text with optional timestamped chunks.
Inputs
- Chute URL — Select a specific speech-to-text chute to use or provide a custom chute URL to target a different model.
- Audio — Audio input for transcription; leave empty to use binary data from a previous node, or provide an audio URL or base64-encoded audio string to override.
- Additional Options — Optional parameters including audio language code (e.g., 'en'), whether to include timestamped chunks in the output, and a maximum timeout in seconds to limit transcription time.
Output shape
a single JSON object containing transcription text; optionally includes an array of timestamped chunks if requested.
If 'Include Chunks' is enabled, the output includes both the continuous transcript text and a 'chunks' array with start/end times and texts. On timeout, the node errors and can trigger retries. Audio input priority is binary data from previous node, then 'Audio' field URL/base64 input.
Examples
Example 1: Transcribing an MP3 audio file from a previous HTTP Request node's binary data using the default Chutes.ai speech-to-text chute.
Set resource to 'Speech-to-Text', operation to 'Transcribe', leave 'Audio' empty to use previous binary data, optionally set language to 'en' and disable 'Include Chunks' for a simple transcript.
Example 2: Using a custom Chutes.ai chute URL for transcription with an audio file URL override and including timestamped chunks.
Set resource to 'Speech-to-Text', operation to 'Transcribe', specify the custom chute URL in 'Chute', enter the audio file URL in 'Audio', set 'Include Chunks' to true in additional options to receive detailed timings along with the transcript.