Nova 3 General

This documentation is valid for the following list of our models:

• nova-3-general

Try in Playground

Nova-3 General — multilingual speech-to-text model from Deepgram with automatic language detection across 30+ languages, optimized for real-time and batch transcription with speaker diarization and advanced audio intelligence features.

This model uses per-minute billing. The cost of audio transcription is based on the duration of the input audio file.

Setup your API Key

If you don't have an API key for the AI/ML API yet, feel free to use our Quickstart guide.

API Schemas

post

Body

application/json

modelstring · enumRequiredPossible values:

urlstring · uriOptional

URL of the input audio file. Provide either url or audio — exactly one is required, not both.

Example: https://audio-samples.github.io/samples/mp3/blizzard_primed/sample-0.mp3

audiostring · binaryOptional

The audio file to transcribe. Provide either url or audio — exactly one is required, not both.

languagestringOptional

The BCP-47 language tag that hints at the primary spoken language. Depending on the model and API endpoint you choose, only certain languages are available.

Example: en

punctuateboolean · nullableOptional

Adds punctuation and capitalization to the transcript.

smart_formatboolean · nullableOptional

Applies formatting to transcript output. When set to true, additional formatting will be applied to transcripts to improve readability.

diarizeboolean · nullableOptional

Recognizes speaker changes. Each word in the transcript will be assigned a speaker number starting at 0.

multichannelboolean · nullableOptional

Enable Multichannel transcription, can be true or false.

paragraphsboolean · nullableOptional

Splits audio into paragraphs to improve transcript readability.

utterancesboolean · nullableOptional

Segments speech into meaningful semantic units.

utt_splitnumberOptional

Seconds to wait before detecting a pause between words in submitted audio.

detect_languageany ofOptional

Enables language detection to identify the dominant language spoken in the submitted audio.

string[]Optional

or

boolean · nullableOptional

detect_entitiesboolean · nullableOptional

When Entity Detection is enabled, the Punctuation feature will be enabled by default.

intentsboolean · nullableOptional

Recognizes speaker intent throughout a transcript or text.

sentimentboolean · nullableOptional

Recognizes the sentiment throughout a transcript or text.

topicsboolean · nullableOptional

Detects topics throughout a transcript or text.

summarizeany ofOptional

Summarizes content. For Listen API, supports string version option. For Read API, accepts boolean only.

stringOptional

or

booleanOptional

custom_intentany ofOptional

A custom intent you want the model to detect within your input audio if present. Submit up to 100.

stringOptional

or

string[]Optional

custom_intent_modestring · enumOptional

Sets how the model will interpret strings submitted to the custom_intent param. When strict, the model will only return intents submitted using the custom_intent param. When extended, the model will return its own detected intents in addition to those submitted.

Possible values:

custom_topicany ofOptional

A custom topic you want the model to detect within your input audio if present. Submit up to 100.

stringOptional

or

string[]Optional

custom_topic_modestring · enumOptional

Sets how the model will interpret strings submitted to the custom_topic param. When strict, the model will only return topics submitted using the custom_topic param. When extended, the model will return its own detected topics in addition to those submitted.

Possible values:

keytermany ofOptional

Keyterm Prompting (Nova-3 General only): boost recognition of specialized terms, brands, and proper nouns. Pass a single term or an array of terms.

stringOptional

or

string[]Optional

filler_wordsboolean · nullableOptional

Filler Words can help transcribe interruptions in your audio, like "uh" and "um".

profanity_filterboolean · nullableOptional

Profanity Filter looks for recognized profanity and converts it to the nearest recognized non-profane word or removes it from the transcript completely.

redactany ofOptional

Redact sensitive information from the transcript. Common values: pii, pci, numbers.

stringOptional

or

string[]Optional

replaceany ofOptional

Search for terms and replace them in the transcript. Provide "find:replace" pairs.

stringOptional

or

string[]Optional

searchany ofOptional

Search for terms or phrases in submitted audio.

stringOptional

or

string[]Optional

dictationboolean · nullableOptional

Identifies and extracts key entities from content in submitted audio.

measurementsboolean · nullableOptional

Spoken measurements will be converted to their corresponding abbreviations.

numeralsboolean · nullableOptional

Numerals converts numbers from written format to numerical format.

encodingstring · enumOptional

Expected encoding of the submitted audio (used mainly for raw audio).

Possible values:

tagany ofOptional

Labels your requests for the purpose of identification during usage reporting.

stringOptional

or

string[]Optional

extraany ofOptional

Arbitrary key-value pairs that are attached to the API response for usage in downstream processing.

stringOptional

or

string[]Optional

mip_opt_outboolean · nullableOptional

Opt out of the Deepgram Model Improvement Program for this request.

versionstringOptional

Model version to use (e.g. "latest" or a specific version string). Defaults to latest.

Responses

200Success

application/json

generation_idstringRequired

The unique identifier of the created transcription task. Use this ID to retrieve the result via GET /v1/stt/{generation_id}.

Example: a1b2c3d4-e5f6-7890-abcd-ef1234567890

post/v1/stt/create

cURLJavaScriptPython

curl -L \
  --request POST \
  --url 'https://api.aimlapi.com/v1/stt/create' \
  --header 'Authorization: Bearer <YOUR_AIMLAPI_KEY>' \
  --header 'Content-Type: application/json' \
  --data '{
      "model": "nova-3-general",
      "url": "https://audio-samples.github.io/samples/mp3/blizzard_primed/sample-0.mp3"
    }'

200Success

{
  "generation_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890"
}

Quick Example: Processing a Speech Audio File via URL

Let's transcribe the following audio fragment:

import time
import requests
import json

base_url = "https://api.aimlapi.com/v1"
# Insert your AIML API Key instead of <YOUR_AIMLAPI_KEY>:
api_key = "<YOUR_AIMLAPI_KEY>"

# Creating and sending a speech-to-text conversion task to the server
def create_stt():
    url = f"{base_url}/stt/create"
    headers = {
        "Authorization": f"Bearer {api_key}",
    }

    data = {
        "model": "nova-3-general",
        "url": "https://audio-samples.github.io/samples/mp3/blizzard_primed/sample-0.mp3"
    }

    response = requests.post(url, json=data, headers=headers)

    if response.status_code >= 400:
        print(f"Error: {response.status_code} - {response.text}")
    else:
        response_data = response.json()
        print(response_data)
        return response_data

# Requesting the result of the task from the server using the generation_id
def get_stt(gen_id):
    url = f"{base_url}/stt/{gen_id}"
    headers = {
        "Authorization": f"Bearer {api_key}",
    }
    response = requests.get(url, headers=headers)
    return response.json()

# First, start the generation, then repeatedly request the result from the server every 10 seconds.
def main():
    stt_response = create_stt()
    gen_id = stt_response.get("generation_id")

    if gen_id:
        start_time = time.time()
        timeout = 600
        while time.time() - start_time < timeout:
            response_data = get_stt(gen_id)

            if response_data is None:
                print("Error: No response from API")
                break

            status = response_data.get("status")

            if status == "waiting" or status == "active":
                print("Still waiting... Checking again in 10 seconds.")
                time.sleep(10)
            else:
                print("Processing complete:\n", response_data["result"]["results"]["channels"][0]["alternatives"][0]["transcript"])

                # Uncomment the line below to print the entire result object with all service data
                # print("Processing complete:\n", json.dumps(response_data["result"], indent=2, ensure_ascii=False))
                return response_data

        print("Timeout reached. Stopping.")
        return None


if __name__ == "__main__":
    main()

Response

{'generation_id': 'b2c3d4e5-f6a7-8901-bcde-f12345678901'}
Still waiting... Checking again in 10 seconds.
Processing complete:
{
  "status": "completed",
  "result": {
    "metadata": {
      "request_id": "b2c3d4e5-f6a7-8901-bcde-f12345678901",
      "created": "2025-06-04T12:00:00.000Z",
      "duration": 11.4,
      "channels": 1,
      "models": ["nova-3-general"],
      "model_info": {
        "nova-3-general": {
          "name": "nova-3-general",
          "version": "2024-12-18.0",
          "arch": "nova-3"
        }
      }
    },
    "results": {
      "channels": [
        {
          "alternatives": [
            {
              "transcript": "He doesn't belong to you, and I don't see how you have anything to do with what is be his power, if he possess only that from this stage to you.",
              "confidence": 0.9921875,
              "words": [
                {
                  "word": "he",
                  "start": 0.32,
                  "end": 0.4,
                  "confidence": 0.9921875,
                  "speaker": 0,
                  "punctuated_word": "He"
                }
              ]
            }
          ]
        }
      ]
    }
  }
}