safe_text 2.0.0

A Flutter package for filtering out bad words from text inputs and detecting phone numbers in various formats including multiplier words.

Safe Text is a high-performance Flutter package designed to filter out offensive language (profanity) and detect phone numbers. Version 2.0.0 introduces a state-of-the-art Aho-Corasick engine for near-instant filtering across multiple languages.

🚀 What's New in 2.0.0

• Aho-Corasick Algorithm: Near-instant multi-pattern search (O(N) complexity).
• Extreme Speed: Up to 20x faster than legacy regex loops.
• Multilingual Excellence: Support for 75+ languages using full human-readable names.
• Modular API: Separate trackers for profanity (SafeTextFilter) and phone numbers (PhoneNumberChecker).
• Memory Efficient: Single-pass string building using StringBuffer.

Features

• Blazing Fast Profanity Filtering: Scans thousands of bad words in a single pass of the text.
• Advanced Normalization: Catches common bypasses like f@ck or b4dass with zero overhead.
• Comprehensive Phone Detection: Detects numbers in digits, words, mixed formats, and repeats (e.g., "triple five").
• Customizable: Add your own extra words or exclude specific phrases.
• Non-blocking: Phone detection and heavy filtering can run in separate isolates.

Getting started

Add safe_text to your pubspec.yaml:

dependencies:
  safe_text: ^2.0.0

1. Profanity Filtering (SafeTextFilter)

The new SafeTextFilter class is the high-performance replacement for the legacy SafeText logic.

Initialization

Initialize with your desired language (defaults to English) to build the search Trie:

import 'package:safe_text/safe_text.dart';

void main() async {
  // Option 1: Initialize with a specific language
  await SafeTextFilter.init(language: Language.english);

  // Option 2: Initialize with a custom list of languages
  await SafeTextFilter.init(languages: [Language.english, Language.hindi, Language.spanish]);

  // Option 3: Initialize with ALL supported languages
  await SafeTextFilter.init(language: Language.all);
}

Filtering Text

String filtered = SafeTextFilter.filterText(
  text: "Hello b4dass!",
  extraWords: ["example"], // Optional
  excludedWords: ["friend"], // Optional
  useDefaultWords: true,
  fullMode: true,
  obscureSymbol: "*",
);
// Result: "Hello ******!"

Checking for Bad Words

bool hasBadWord = await SafeTextFilter.containsBadWord(
  text: "Don't be a pendejo",
);

2. Phone Number Detection (PhoneNumberChecker)

bool hasPhone = await PhoneNumberChecker.containsPhoneNumber(
  text: "Call me at nine 7 eight 3 triple four",
  minLength: 7,
  maxLength: 15,
);

3. Legacy Support (SafeText)

The original SafeText class is still available but marked as @Deprecated. It internally redirects to the new modular classes. We recommend migrating to the new API for a better developer experience.

Parameters

Parameter	Type	Default	Description
text	String	Required	The input string to process.
extraWords	List<String>?	null	Add custom patterns to the filter list.
excludedWords	List<String>?	null	Phrases that should NOT be filtered.
useDefaultWords	bool	true	Include the built-in language patterns.
fullMode	bool	true	true: ****, false: f**k.
obscureSymbol	String	*	Symbol used for obscuring.

Why is v2.0.0 so fast?

Legacy versions used a nested loop approach (for every bad word, run a regex). With 10,000+ words, this grew exponentially slow.

Aho-Corasick builds a Finite State Automaton (Trie) from the word list. The engine then scans your text exactly once, matching all possible patterns simultaneously in O(N) time where N is the length of your text.

Contributing

Pull requests are welcome. For major changes, please open an issue first.

Data Source

SafeText uses the comprehensive List of Dirty, Naughty, Obscene, and Otherwise Bad Words repository. This dataset includes:

• 75+ Dialects/Languages
• 55,000+ curated words

We are grateful to the contributors of this dataset for providing a robust foundation for profanity filtering.

Authors

_{Ronit Rameja}

Connect on LinkedIn: Ronit Rameja