All products

Connect Cartesia speech APIs with AI

Add speech synthesis or transcription to an application using a real voice ID, the correct API version and a server-side credential for the chosen workflow.

To connect Cartesia speech APIs with AI, start with the direction the information needs to travel. Reading a product introduction aloud and turning a recording into text are different jobs. The cartesia-api skill helps choose the appropriate endpoint, client and delivery method for each.

A streaming interface may need to play audio as it arrives; a simple export may only need the completed output. The workflow checks the API version and available library voices before asking the service to generate anything. Recording or reproducing another person’s voice requires the appropriate rights and explicit consent.

What it does

  • Choose between speech synthesis and transcription for an application.
  • Use a real, authorised library voice rather than inventing an identifier.
  • Select an HTTP or streaming connection to match the experience.
  • Keep API credentials on the server instead of exposing them to visitors.
  • Check the version and language settings before testing a short sample.

Install and use

  1. Install the skill and use an authorised Cartesia account, an API key and a server-side SDK or HTTPS client. The optional hosted MCP uses a separately authorised connection.
  2. Ask, “Read this short product introduction with an available catalogue voice. Verify its ID and give me a server-side integration for our app.”
  3. Listen to the sample or inspect the transcript before moving the settings into the main application.

FAQ

Source

Made by Cartesia. Original skill: cartesia-api, published on GitHub under the MIT licence.

View on GitHub