Open-source ML-powered profanity filter with TensorFlow.js toxicity detection, leetspeak & Unicode obfuscation resistance. 21M+ ops/sec, 23 languages, React hooks, LRU caching. npm & PyPI.

Self-hosted content moderation API that outperforms Amazon Comprehend. 100% offline, your data never leaves your server. Text + Image moderation.

This is a simple python program which uses a machine learning model to detect toxicity in tweets, developed in Flask.

This demo shows the functionality of the Voximplant instant messaging SDK, including silent supervision by a bot.

AntiToxicBot is a bot that detects toxics in a chat using Data Science and Machine Learning technologies. The bot will warn admins about toxic users. Also, the admin can allow the bot to ban toxics.

A supervised learning based tool to identify toxic code review comments

🛡️ Programmable Guardrails for LLM Applications in Java. A framework-agnostic toolkit for input/output validation, PII masking, and jailbreak detection. The Java alternative to NVIDIA NeMo Guardrails.

This is a simple python program which uses a machine learning model to detect toxicity in tweets, GUI in Tkinter.

NLP deep learning model for multilingual toxicity detection in text 📚

Enterprise LLM Evaluation & Responsible AI Framework — Benchmark bias, hallucination, PII leakage, and toxicity across Healthcare, BFSI, Retail & Legal industries. Supports OpenAI, Anthropic, Gemini & HuggingFace. Python SDK + CLI + Web Dashboard. 191 tests. Compliance-ready reports.

Telegram bot that detects toxic comments based on Perspective API

Using Language Models to Identify and Classify Toxicity Inside In-Game Chat

Simple Multi-Language HTTP Server Text Toxicity Detector

It is a trained Deep Learning model to predict different level of toxic comments. Toxicity like threats, obscenity, insults, and identity-based hate.

FastAPI-based multilingual NLP backend for code-mixed text analysis — featuring language detection, sentiment & toxicity analysis, translation, and Indic script conversion.

This repository presents a comprehensive comparison of traditional machine learning, deep learning, encoder-based, and decoder-based large language models (LLMs) for Chinese toxic comment detection.

State-of-the-art Chinese toxicity detection — dual-LoRA ensemble (Qwen3-8B) exceeding PCR-ToxiCN SOTA with homophone attack defense. 中文毒性偵測 · 超越 SOTA · 諧音攻擊防禦

This work focuses on the development of machine learning models, in particular neural networks and SVM, where they can detect toxicity in comments. The topics we will be dealing with: a) Cost-sensitive learning, b) Class imbalance

An Explainable Toxicity detector for code review comments. Published in ESEM'2023

Benchmarking and enhancing French toxicity detection with Chain-of-Thought fine-tuning — a large-scale dataset (53k samples) and SLM-based methodology achieving state-of-the-art results.