This addon let's you add audio from (aws) polly to your deck.

Ankiweb(2090845100)

{
  "access_id": "",
  "access_key": "",
  "audio_field": "audio",
  "engine": "standard",
  "query_fields": ["text"],
  "query_fields_seperator": "<break/>",
  "template": "{text}",
  "region": "eu-west-1",
  "voice": ""
}

• access_id: The id (NOT YOUR LOGIN CREDENTIAL) of the aws account you want to use.
• access_key: The key (NOT THE PASSWORD) for the aws account you want to use.
• audio_field: The name of the audio field you want to save the audio reference to.
• engine: The name of the engine you want to use (standard or neural).
• query_fields: A list of fields you want the audio from.
• query_fields_seperator: The seperator that will be used to combine those fields.
• template: The template you can use to supply some ssml decorators. You can also use {combined} or single field names as a reference like {field1}, {field2}. The have to be in the query_fields list tho.
• region: The region of your aws instance.
• voice: The aws polly voice you want to use.

If I would want to create a audio file from the fields singular1 and singular2, which contain the first and second conjugated forms of some german verb, seperated by a short break of 1s and save the result to the audio field, I would do it like this:

{
  "access_id": "idididididididid",
  "access_key": "keykeykeykeykey",
  "audio_field": "audio",
  "engine": "standard",
  "query_fields": ["singular1", "singular2"],
  "query_fields_seperator": "<break time=\"1s\">",
  "template": "{combined}",
  "region": "eu-west-1",
  "voice": "Hans"
}

you could also change the template to this (same result):

{
"template": "{singular1}<break time=\"1s\">{singular2}"
}

or this (same result):

{
"template": "<speak>{singular1}<break time=\"1s\">{singular2}</speak>"
}

Link to docs