Skip to content

Cartesia Speech to Text

Cartesia Speech to Text turns recorded audio files into text with Cartesia’s Ink Whisper model. Unlike other services, Cartesia does not detect the language: it transcribes in the one you set, Spanish by default. It returns one Data per file, with the transcript in the text field.

You need a Cartesia API credential with your Cartesia account key. See Credentials. The audio is sent to Cartesia to be transcribed.

This is not the voice node for calls: to listen to a caller during a live call, the Voice Agent uses the Cartesia STT node. This component waits for the whole file.

When to use it

  • Turn voice notes or recordings into text when you already use Cartesia for voices and know the language spoken.
  • Let an agent transcribe audio when it needs to, using the component as a tool.
  • Transcribe a list of audio files in one go with Batch Mode.

If you don’t know which language the audio will be in, or several are mixed, use one that detects it, such as Deepgram Speech to Text or Groq Whisper Speech to Text.

Configuration

FieldWhat it does
API Key (required)Your Cartesia key. The key is filled in for you when you choose your Cartesia credential.
Audio (required)The audio (or video) file to transcribe: WAV, MP3, OGG, M4A, WEBM… Connect it to the component that brings the file, for example a webhook, a chat with attachments or storage. It accepts several files.
LanguageThe language of the audio, as its code, for example es or en. Default: es. Cartesia does not detect it: if you leave it empty, it transcribes as if it were English.

Advanced options

FieldWhat it does
ModelCartesia’s transcription model. Default: ink-whisper, the only one available for files.

Output

The Data output returns one record per file:

{
"text": "Hola, buenos días. Su cita está confirmada para el martes a las 10 y media.",
"language": "es",
"file_name": "voice_note.ogg",
"provider": "Cartesia",
"model": "ink-whisper",
"duration": 5.1
}

duration is the length of the audio in seconds.

Example

An English academy asks its students to send a recording of themselves reading a text. The flow passes each audio to Cartesia Speech to Text with Language set to en, and an agent compares the transcript with the original text to point out the words the student skipped.

Tips

  • Check Language when the audio is not in Spanish: with the wrong value, the transcript comes out wrong.
  • Cartesia writes numbers and times as digits: “a las diez y media” comes out as “a las 10 y media” or “a las 10:30”.
  • Measured in our test environment on 3 November 2026: the sentence “Hola, buenos días. Su cita está confirmada para el martes a las diez y media.” was transcribed in 0.2 s, word for word.