ai_sensitive_content_classifier

A Dart/Flutter package that classifies text and images for sensitive content using the Google Generative AI (Gemini) API.

Supports input as plain text, Uint8List, ui.Image, or Flutter ImageProvider.

🚀 Features

• 🔍 Classifies content(text/images or bytes directly) into types like gore, violence, nudity, racism, hateSpeech, offensive, or notSensitive
• 🧠 Powered by Google Gemini (gemini-2.0-flash-lite by default)
• 🖼️ Supports direct analysis from Flutter UI images (📸 Support for ui.Image, Uint8List, or ImageProvider (e.g., AssetImage, NetworkImage))
• 🧪 Easily customizable model config (temperature, topP, maxOutputTokens, etc.)
• ✅ JSON schema validation for safer AI response handling
• 🧠 JSON schema validation for structured AI responses
• 🛡 No content filtering: all Gemini safety filters disabled to ensure full sensitivity analysis

🛠 Installation

Add this to your pubspec.yaml:

dependencies:
  ai_sensitive_content_classifier: ^0.1.3

Then run

flutter pub get

Usage:

import 'package:ai_sensitive_content_classifier/ai_sensitive_content_classifier.dart';

final classifier = AiSensitiveContentDetector(
  apiKey: 'your-gemini-api-key',
);

final result = await classifier.analyseIsSensitiveContent(
  text: 'This is a violent message',
);

print(result?.isSensitive); // true/false
print(result?.textClassification); // e.g., "violence"

Configurations:

AiSensitiveContentDetector({
  required String apiKey,
  String model = 'gemini-2.0-flash-lite',
  double temperature = 0.1,
  double topP = 0.95,
  int topK = 64,
  int maxOutputTokens = 8192,
})