Profanity Filter

Profanity filter detects and censors profane words and phrases as audio is being transcribed. When profane words and phrases are detected, all letters except the first are masked (e.g. f***). The original words can still be retrieved using the Word.orig_text field.

The profanity filter can be enabled by setting the enable_profanity_filter to true. By default, the profanity filter is disabled. The profanity filter supports both real-time and asynchronous use cases.

Example

In this example, we transcribe a short audio file with profanity filter enabled.

Python Lib

profanity_filter.py

from soniox.transcribe_file import transcribe_file_short
from soniox.speech_service import SpeechClient, set_api_key

set_api_key("<YOUR-API-KEY>")


def main():
    with SpeechClient() as client:
        result = transcribe_file_short(
            "../test_data/test_audio_profanity.mp3",
            client,
            enable_profanity_filter=True,
        )
        print("Words: " + " ".join(w.text for w in result.words))


if __name__ == "__main__":
    main()

Run

python3 profanity_filter.py

Output

This is f****** great . No b******* whatsoever

Node Lib

profanity_filter.js

const { SpeechClient } = require("@soniox/soniox-node");

// Do not forget to set your Soniox API key.
const speechClient = new SpeechClient();

(async function () {
    const result = await speechClient.transcribeFileShort(
        "../test_data/test_audio_profanity.mp3",
        {
            enable_profanity_filter: true,
        }
    );

    console.log(`Words: ${result.words.map((word) => word.text).join(" ")}`);
})();

Run

node profanity_filter.js

Output

This is f****** great . No b******* whatsoever