code-detection

Unified solution for SemEval-2026 Task 13: GenAI Code Detection & Attribution. A modular framework covering Subtasks A (Detection), B (Authorship), and C (Mixed-Source Analysis), designed to distinguish and analyze LLM-generated vs. Human code with a centralized pipeline.

A tiny byte-level multi-head content classifier (~1.5M params, ~200KB ONNX, <6ms). Classifies code, text, markup, config, images, binary, secrets, 62 code languages, 30 text languages, 90 MIME types from raw bytes — no tokenizer needed.

Advanced machine learning system for detecting whether code was written by artificial intelligence or humans. Features intelligent ensemble models, GitHub repository analysis, and comprehensive explainability with smart contradiction detection.

Codify - Auto detect & highlight & compile raw text source codes in web

Detect and filter auto-generated Go code — fast, zero-config, table-driven. Supports sqlc, protobuf, templ, mockgen, wire, stringer, and more.

CodeCheck is a python app that scans an entire project’s directory and codebase to detect what files and % of it are vibecoded.

Program to detect programming language from the a code snippet

AI-Powered Code Analyzer AI-Powered Code Analyzer is an intelligent developer-focused application designed to automate and simplify the process of analyzing source code, identifying potential issues, and providing actionable feedback for improving code quality. Built with Python and Streamlit, the project combines static code analysis, abstract

PHP library for automatic programming language detection

Codetector is a modular AI-generated content (AIGC) detection framework that allows easy integration of custom datasets, large language models (LLMs), and detection methods. With features like crash recovery and unit-tested code, it provides a flexible and robust pipeline for AIGC generation and detection.