Tagging TTS Requests

tag string

Text to Speech Request Text to Speech Stream All available languages

Deepgram’s Tagging feature allows you to label your Text-to-Speech API requests for the purpose of identification during usage reporting. You can also apply tags to API Keys; if you do, any tags applied to the API Key running the API request will also be applied to the request itself.

To enable Tagging, when you call Deepgram’s Text-to-Speech API, add a tag parameter in the query string and set it to the tag you would like to recognize:

tag=VALUE

To generate speech from text, run the following cURL command in a terminal or your favorite API client. Please be aware that once you have set a tag, you cannot modify it.

curl \
  --request POST \
  --header 'Authorization: Token YOUR_DEEPGRAM_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{"text": "Hello, how can I help you today?"}' \
  --url 'https://api.deepgram.com/v1/speak?model=aura-2-thalia-en&tag=VALUE'\
  --output test_audio.wav

Once applied, you can identify tags associated with API requests returned by the Get All Requests, Get Request, and Get Fields endpoints.

You can also directly query requests by tag at the Summarize Usage endpoint.

curl \
  --request GET \
  --header 'Authorization: Token YOUR_DEEPGRAM_API_KEY' \
  --header 'content-type: application/json' \
  --url 'https://api.deepgram.com/v1/projects/PROJECT_ID/usage?tag=TEST'

Tags are limited to 128 characters per tag and 500 unique tags per day.

If your tag or extra metadata includes spaces or special characters, be sure to URL encode it:

tag=marketing%20team or tag=marketing+team

To apply multiple tags or multiple extra key-value pairs, submit the query parameter multiple times in your API request:

tag=marketing&tag=legal

Extra Metadata is a similar feature to Tagging. Where Tagging is primarily intended for tracking and filtering usage, Extra Metadata is useful for passing data to downstream processing steps.

Below is a comparison table summarizing the main differences between the two features: