Use FastPII with LangChain to anonymize or preprocess text before it reaches models and external services.

LangChain integration

FastPII provides LangChain-friendly wrappers for anonymization, preprocessing, and a reusable tool for filtering PII from text.

Installation

pip install fastpii[langchain]

Setup

Available imports:

from fastpii.integrations.langchain import (
    PIIAnonymizer,
    PIIPreprocessor,
    create_pii_filter_tool,
)

Usage examples

Anonymize text

from fastpii.integrations.langchain import PIIAnonymizer

anonymizer = PIIAnonymizer(regions=["cz"])

text = "Jan Novak, rodne cislo 8001011234"
clean_text = anonymizer.anonymize(text)
custom_text = anonymizer.anonymize(text, replacement="[PII]")
same_as_anonymize = anonymizer(text)

Preprocess text with different actions

from fastpii.integrations.langchain import PIIPreprocessor

preprocessor = PIIPreprocessor(regions=["cz"])

text = "Jan Novak, rodne cislo 8001011234"

redacted = preprocessor.preprocess(text, action="redact")
masked = preprocessor.preprocess(text, action="mask")
removed = preprocessor.preprocess(text, action="remove")
default_result = preprocessor(text)

Create a LangChain tool

from fastpii.integrations.langchain import create_pii_filter_tool

tool = create_pii_filter_tool()

result = tool.run("Jan Novak, rodne cislo 8001011234")