How to Filter Toxic Keywords in Python

Filter toxic keywords from text by replacing each occurrence with asterisks, useful as a basic guardrail for LLM inputs.

Easy Python 3.9+ Aug 9, 2026 AI & LLM integration patterns 12 views 0 copies

guardrails text-filtering llm-safety string-manipulation

Python code

23 lines

Python 3.9+

TOXIC_KEYWORDS = ["insult", "threat", "hate", "violence", "spam"]


def guardrails_filter(text: str, keywords: list[str] | None = None) -> str:
    """Filter out toxic keywords from the given text.

    Args:
        text: The input text to filter.
        keywords: Optional keyword list. Defaults to TOXIC_KEYWORDS.

    Returns:
        The filtered text with toxic keywords replaced by '*' characters.
    """
    keywords = keywords or TOXIC_KEYWORDS
    filtered_text = text
    for keyword in keywords:
        filtered_text = filtered_text.replace(keyword, "*" * len(keyword))
    return filtered_text


if __name__ == "__main__":
    sample_text = "This is a hate message and a threat to violence."
    print(guardrails_filter(sample_text))

Output

stdout

This is a **** message and a ****** to ********.

How it works

The function iterates through a list of keywords and uses str.replace() to substitute each toxic word with asterisks of the same length. The optional keywords parameter defaults to a predefined list, letting you reuse the filter with custom lists. Because replace() handles substring matches, this approach works for any text where toxic terms appear as standalone or embedded words.

Common mistakes

Replacing keywords with fixed-length asterisks instead of matching keyword length
Forgetting that substring matching may filter non-toxic words containing the keyword
Not handling case sensitivity — 'Hate' won't be caught by 'hate'

Variations

Use regex with word boundaries to avoid partial-word matches
Convert text and keywords to lowercase before filtering for case-insensitive matching

Real-world use cases

Pre-processing user prompts before sending to an LLM API to block harmful or spammy content.
Sanitizing chat messages in a moderation queue to strip offensive terms before display.
Redacting sensitive or banned words in logs or analytics pipelines before storage.

How to Filter Toxic Keywords in Python

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from AI & LLM integration patterns

Tutorials

Quizzes

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from AI & LLM integration patterns

Keep learning

Tutorials

Quizzes