AI Trending

Project directory

Projects

Open-source AI projects tracked by the leaderboard, ordered by GitHub stars.

Python 1,812 stars

NotJoeMartinez/yt-fts

YouTube Full Text Search - Search all of YouTube from the command line

### Overview: `yt-fts` is a command-line tool designed to search YouTube channels for specific keywords or phrases. It utilizes yt-dlp to scrape subtitles from YouTube channels and stores them in a SQLite database for easy querying. The project also supports semantic search using embeddings from OpenAI, Gemini, and Chroma, making it a powerful tool for finding relevant YouTube content based on keywords and semantic similarity. ### Key Features: - **Command Line Interface (CLI):** The primary interface for interacting with the tool. - **Subtitle Scrape:** Downloads subtitles for YouTube channels and playlists. - **Semantic Search:** Supports semantic search using embeddings from OpenAI, Gemini, and Chroma. - **Database Management:** Manages a SQLite database containing the scraped subtitles. - **Parallel Downloads:** Allows for parallel downloads of subtitles to speed up the process. - **Customization:** Options for specifying language, number of jobs, and using browser cookies for authentication. - **List and Update Functions:** Provides commands to list saved channels, videos, and transcripts, as well as to update subtitles for specific channels. ### Tech Stack: Python ### Why It Matters: `yt-fts` is a valuable tool for developers and content creators who need to quickly find relevant YouTube content. Its ability to perform semantic search and manage subtitles in a database makes it particularly useful for applications that require advanced search functionalities. The project's command-line interface and support for multiple search engines make it accessible and powerful for a wide range of users. Repository: NotJoeMartinez/yt-fts Primary Language: Python
Python 1,810 stars

pemistahl/lingua-py

The most accurate natural language detection library for Python, suitable for short text and mixed-language text

Overview: Lingua-py is a highly accurate natural language detection library for Python, designed for short text and mixed-language text. It is suitable for preprocessing linguistic data in applications like text classification and spell checking. Key features: - Accurate language detection for short text and mixed-language text - Supports multiple languages - Fast performance and low memory usage - Offline functionality Tech stack: Python Why it matters: Lingua-py is essential for natural language processing tasks, offering a lightweight solution for language detection that works well with short text and mixed-language scenarios.
Python 1,810 stars

zi-yue-1129/DATAGEN

DATAGEN: AI-driven multi-agent research assistant automating hypothesis generation, data analysis, and report writing.

Overview: DATAGEN is an AI-driven platform that automates hypothesis generation, data analysis, and report writing through a multi-agent system, leveraging LangChain, OpenAI's GPT models, and LangGraph. Key features: - Intelligent Analysis Core: Advanced hypothesis engine for AI-driven research direction optimization. - Enterprise Data Processing: Robust data cleaning and transformation, scalable analysis pipelines, and automated quality assurance. - Dynamic Visualization Suite: Interactive data visualization, custom report generation, and automated insight extraction. Tech stack: Python. Why it matters: DATAGEN revolutionizes data analysis by combining an advanced multi-agent system with intelligent automation, providing enterprise-grade performance and seamless integration across various analysis phases. Repository: zi-yue-1129/DATAGEN Primary language: Python GitHub stars: 1809 GitHub description: DATAGEN: AI-driven multi-agent research assistant automating hypothesis generation, data analysis, and report writing.
Jupyter Notebook 1,809 stars

a2aproject/a2a-samples

Samples using the Agent2Agent (A2A) Protocol

Overview: This repository contains samples demonstrating the use of the Agent2Agent (A2A) Protocol, showcasing how to build and interact with multi-agent systems across various programming languages and environments. Key features: - Multi-Agent Communication - Python CLI Host - Jupyter Notebook Integration - Firebase Studio Integration Tech stack: Jupyter Notebook Why it matters: This repository provides a practical, accessible way to learn and experiment with multi-agent systems, leveraging the A2A protocol to facilitate secure and interoperable communication between agents.
Python 1,809 stars

undertheseanlp/underthesea

Underthesea - AI Assistant

### Overview: Underthesea is an open-source Agentic AI Toolkit that provides Vietnamese Natural Language Processing capabilities. It supports multiple AI providers and includes Python modules for efficient and effective text processing. ### Key Features: - **Multi-Provider Support**: Integrates with popular AI providers like OpenAI, Azure OpenAI, Anthropic Claude, and Google Gemini. - **Built-in Vietnamese NLP**: Offers a suite of Python modules for Vietnamese NLP tasks. - **Zero External Dependencies**: Communicates with LLM APIs using Python's standard library, eliminating the need for external packages. - **Agent Support**: Provides a multi-provider AI Agent that can be configured with different providers. ### Tech Stack: Python ### Why It Matters: Underthesea is a powerful tool for anyone looking to integrate AI capabilities into their projects using Python. Its flexibility and ease of use make it an excellent choice for developers and researchers alike.
Python 1,808 stars

Trusted-AI/AIX360

Interpretability and explainability of data and machine learning models

Overview: AI Explainability 360 is an open-source library designed to enhance the interpretability and explainability of datasets and machine learning models, particularly focusing on datasets and text-based models. Key features: - Support for various data types including tabular, text, images, and time series data. - Comprehensive set of algorithms for different dimensions of explanations. - Proxy explainability metrics for improved understanding. - Interactive experience for beginners and advanced users. - Extensive documentation and tutorials for both data scientists and developers. - Extensive community support and contributions encouraged. Tech stack: Python Why it matters: AI Explainability 360 provides a valuable tool for making complex machine learning models more understandable, crucial for ensuring trust and reliability in AI applications.
Python 1,808 stars

google/budoux

### Overview: BudouX is a language-neutral, standalone, and small machine learning-powered line break organizer. It supports various languages and can be used with or without dependencies. ### Key features: - **Standalone**: Works without external word segmenters. - **Small**: Only 15 KB in size. - **Language-neutral**: Supports any language by training a model. - **HTML Input**: Handles HTML inputs to ensure phrases are not split across lines. ### Tech stack: Python ### Why it matters: BudouX simplifies the process of managing line breaks in text by leveraging machine learning, making it a valuable tool for developers and content creators.
Jupyter Notebook 1,808 stars

sachinruk/deepschool.io

Deep Learning tutorials in jupyter notebooks.

Overview: DeepSchool.io is a comprehensive repository of Jupyter Notebooks designed to simplify the learning of Deep Learning. It includes tutorials on data science, deep learning, and Bayesian learning, making it accessible and practical for developers. Key features: - Data Science tutorials using Pandas - Deep Learning with Keras - Bayesian Learning with PyMC3 Tech stack: Jupyter Notebook Why it matters: This initiative aims to make Deep Learning more accessible, practical, and fun, fostering a community of learners and developers. Repository: sachinruk/deepschool.io GitHub stars: 1808 GitHub description: Deep Learning tutorials in Jupyter notebooks.
Python 1,808 stars

yongzhuo/Keras-TextClassification

中文长文本分类、短句子分类、多标签分类、两句子相似度（Chinese Text Classification of Keras NLP, multi-label classify, or sentence classify, long or short），字词句向量嵌入层（embeddings）和网络层（graph）构建基类，FastText，TextCNN，CharCNN，TextRNN, RCNN, DCNN, DPCNN, VDCNN, CRNN, Bert, Xlnet, Albert, Attention, DeepMoji, HAN, 胶囊网络-CapsuleNet, Transformer-encode, Seq2seq, SWEM, LEAM, TextGCN

### Overview: Keras-TextClassification是一个基于Keras的文本分类框架，支持中文长文本分类、短句子分类、多标签分类以及两句子相似度计算。该框架提供了多种预训练模型，包括FastText、TextCNN、CharCNN、TextRNN、TextRCNN、TextDCNN、TextDPCNN、TextVDCNN、TextCRNN、DeepMoji、SelfAttention、HAN、CapsuleNet、Transformer-encode、SWEM、LEAM、TextGCN等。此外，它还支持自定义模型的训练和预测，如BERT、Word2Vec、随机向量等。 ### Key features: - 支持中文长文本分类 - 支持短句子分类 - 支持多标签分类 - 支持两句子相似度计算 - 提供多种预训练模型，如FastText、TextCNN、CharCNN、TextRNN、TextRCNN、TextDCNN、TextDPCNN、TextVDCNN、TextCRNN、DeepMoji、SelfAttention、HAN、CapsuleNet、Transformer-encode、SWEM、LEAM、TextGCN - 支持自定义模型的训练和预测，如BERT、Word2Vec、随机向量等 ### Tech stack: - Python ### Why it matters: 该框架为自然语言处理领域的开发人员提供了一个强大的工具箱，适用于各种文本分类任务，包括长文本分类、短句子分类、多标签分类以及两句子相似度
Python 1,807 stars

ai-infra-curriculum/ai-infra-engineer-learning

AI Infrastructure Engineer Learning Track - Production ML infrastructure curriculum (2-4 years experience)

### Overview: This repository provides a comprehensive and production-ready learning path for becoming an AI Infrastructure Engineer. It covers building ML infrastructure, deploying production ML systems, implementing end-to-end MLOps pipelines, and deploying cutting-edge infrastructure for LLMs and GPU clusters. The curriculum is aligned with industry requirements and includes hands-on projects and production-ready code stubs with TODO comments. ### Key Features: - ✅ Build ML Infrastructure from scratch (Docker, Kubernetes, cloud platforms) - ✅ Deploy Production ML Systems with auto-scaling and comprehensive monitoring - ✅ Implement End-to-End MLOps pipelines (Airflow, MLflow, DVC) - ✅ Deploy Cutting-Edge LLM Infrastructure (vLLM, RAG, vector databases) - ✅ Scale Training with distributed systems and GPU clusters - ✅ Monitor and Troubleshoot complex ML systems in production - ✅ Optimize Costs across cloud providers (60-80% savings possible) ### Tech Stack: Python ### Why It Matters: This learning path is directly mapped to $120k-$180k AI Infrastructure Engineer roles, providing a practical and industry-aligned approach to mastering AI infrastructure engineering.
HTML 1,806 stars

didi/ChineseNLP

Datasets, SOTA results of every fields of Chinese NLP

Overview: This repository contains shared tasks, datasets, and state-of-the-art results for Chinese Natural Language Processing (NLP). Key features: - Shared tasks and datasets for Chinese NLP - Triangular MT shared task for improving Russian-to-Chinese machine translation - Progress tracking in English and other languages through [NLP-progress](http://nlpprogress.com/) Tech stack: HTML Why it matters: This repository serves as a resource for researchers and practitioners in Chinese NLP, providing essential datasets and state-of-the-art results to advance the field. Repository: didi/ChineseNLP GitHub stars: 1806
Go 1,806 stars

trhhubsd/ARTEX

ARTEX · AI autonomous penetration testing system (multi-agent LLM, Go + Next.js) · AI 自主渗透测试系统 · 百度 agent+ 攻防挑战赛冠军

### Overview: ARTEX is an AI autonomous penetration testing system built with Go and Next.js. It features a multi-agent system for autonomous penetration testing in an isolated lab, where planners and workers share an asset graph and an exploration graph with human-in-the-loop approval and full traffic recording. ### Key Features: - **Multi-Agent System**: Multiple agents collaborate to perform penetration testing tasks. - **Go Backend**: Secure and efficient backend for handling various tasks. - **Next.js Frontend**: User-friendly and responsive frontend for real-time collaboration and data visualization. ### Tech Stack: - **Go**: The primary programming language used for the backend. - **Next.js**: The frontend framework for a seamless user experience. ### Why It Matters: ARTEX is significant because it combines the power of AI with the flexibility of Go, offering a robust solution for autonomous penetration testing. Its ability to handle large-scale and complex penetration testing tasks makes it highly valuable in cybersecurity environments. ### Repository: trhhubsd/ARTEX - **Primary Language**: Go - **GitHub Stars**: 1694 - **Description**: ARTEX is an AI autonomous penetration testing system, featuring a multi-agent system for autonomous penetration testing in an isolated lab. It includes a planner and workers sharing an asset graph and an exploration graph with human-in-the-loop approval and full traffic recording. It ships as a binary or Docker Compose stack. ### Installation: - **Recommended**: Use the provided script to install ARTEX. This includes setting up the database and configuring the LLM. - **Docker Compose**: Alternatively
Python 1,805 stars

OpenAdaptAI/OpenAdapt

For EHRs, payer portals, and any app your team clicks through. A person or an agent shows the task; OpenAdapt turns it into a local program that repeats the steps with no model calls, checks each save, and stops when something doesn't match.

### Overview: OpenAdapt is an open-source, verified execution layer designed for computer-use agents. It compiles tasks into local programs that run without model calls or screenshots, ensuring accurate outcomes and efficient performance. ### Key features: - **Task Compilation**: Converts any task into a local program that runs without model calls or screenshots. - **Model-Free Execution**: Ensures accurate outcomes by reading and confirming records before each save. - **Efficiency**: Reduces model calls and unnecessary steps, improving performance and cost-effectiveness. ### Tech stack: Python ### Why it matters: - **Accuracy and Efficiency**: Ensures tasks are performed accurately and efficiently, reducing the need for costly model calls and unnecessary steps. - **Scalability**: Can be easily integrated into various applications, including EHRs, payer portals, and more. - **User-Friendly**: Designed for agents and people alike, making it accessible for a wide range of users. ### Repository: OpenAdaptAI/OpenAdapt - **GitHub Stars**: 1805 - **Primary Language**: Python - **GitHub Description**: A person or an agent shows the task; OpenAdapt turns it into a local program that repeats the steps with no model calls, checks each save, and stops when something doesn't match. ### Measured on OpenEMR: - **Performance Improvement**: Compared to a human agent, OpenAdapt reduced the time from 36.5 seconds to 22.5 seconds for a new-patient referral in OpenEMR. - **Model Cost**: $0
HTML 1,805 stars

standardagents/dmux

A dev agent multiplexer for git worktrees and coding agents.

Overview: dmux is a developer agent multiplexer designed for managing multiple AI coding agents in isolated git worktrees, allowing parallel development and collaboration. Key features: - Worktree isolation for each agent - Agent support for various AI tools like Claude, Codex, OpenCode, etc. - Goal launches and multi-select launches for agent selection - Smart merging and durable terminals - macOS notifications and built-in file browser - Pane visibility controls and multi-project support - Lifecycle hooks for script execution Tech stack: Node.js, tmux, Git, and various AI agents Why it matters: dmux simplifies the complex process of managing multiple AI agents in parallel, making development more efficient and streamlined.
C 1,805 stars

symisc/sod

An Embedded Computer Vision & Machine Learning Library (CPU Optimized & IoT Capable)

### Overview: SOD is an embedded, modern cross-platform computer vision and machine learning software library designed for real-time applications, providing a common infrastructure for computer vision and accelerating machine perception in open-source and commercial products. ### Key features: - **RealNets Architecture**: A brand new architecture specifically designed for SOD, offering advanced real-time object detection. - **State-of-the-Art Deep Neural Networks**: Includes classic and state-of-the-art deep neural networks, pre-trained models, and patent-free algorithms. - **Cross-Platform and Dependency-Free**: Amalgamated (single C file) and highly optimized, supporting major image formats. - **Easy Interpolation with OpenCV**: Interpolatable with OpenCV or any other proprietary API, making it easy to integrate with existing systems. - **Pre-Trained Models**: Available for most architectures, providing ready-to-use solutions. - **CPU Optimized and Real-Time Capable**: Designed for limited computational resources, embedded systems, and IoT devices. - **Simple, Clean, and Easy to Use API**: Easy to use and clean, with clear documentation and examples. ### Tech Stack: C ### Why it Matters: SOD is essential for real-time computer vision applications, providing a common infrastructure and accelerating machine perception in open-source and commercial products.
HTML 1,803 stars

salesforce/WikiSQL

A large annotated semantic parsing corpus for developing natural language interfaces.

### Overview: WikiSQL is a large annotated semantic parsing corpus designed for developing natural language interfaces to relational databases. It serves as a valuable resource for researchers and developers in the field of AI and NLP. ### Key Features: - **Semantic Parsing Corpus**: Contains annotated data for parsing natural language queries into SQL queries. - **Crowdsourced Annotation**: Data is crowd-sourced, ensuring high accuracy and diversity. - **Reinforcement Learning**: Used in the Seq2SQL project for generating structured queries from natural language. - **Citation**: Follows the citation provided in the README, ensuring reproducibility and credibility. ### Tech Stack: HTML ### Why it Matters: WikiSQL is crucial for researchers and developers aiming to create more intuitive and effective natural language interfaces for databases. Its annotated data and diverse crowd-sourced approach make it an essential resource in the field of AI and NLP. Repository: salesforce/WikiSQL GitHub stars: 1803
Unknown 1,802 stars

alan-ai/alan-sdk-android

The Self-Coding System for Your App — Alan AI SDK for Android

Overview: The Alan AI SDK for Android enables developers to integrate intelligent features into their Android applications by generating business logic and UI in real-time using a self-coding engine. Key features: - Real-time AI-driven feature generation - Embedding an AI agent into Android apps - Integrates with existing app stack, including user interface, business logic, and data management - Enables accurate, context-aware code generation - Creates a safe and validated environment for integrating AI-driven interfaces into apps Tech stack: Java/Kotlin, Android Why it matters: This SDK allows companies to integrate AI-driven interfaces into their existing apps quickly, enabling adaptive, evolving, and scalable software.
Swift 1,802 stars

xmartlabs/Bender

Easily craft fast Neural Networks on iOS! Use TensorFlow models. Metal under the hood.

Overview: Bender is an abstraction layer over MetalPerformanceShaders designed for iOS developers to easily craft and run neural networks, making it simple to integrate pretrained models and support TensorFlow models. Key features: - Define custom network operators - Load models from TensorFlow - Support for MetalPerformanceShaders - Flexibility in receiving parameters Tech stack: Swift Why it matters: Bender simplifies the process of working with neural networks on iOS, making it easier to integrate pretrained models and support TensorFlow models, thus enhancing the capabilities of iOS developers in AI and machine learning projects.
TypeScript 1,800 stars

lmstudio-ai/lmstudio-js

LM Studio TypeScript SDK

Overview: LM Studio TypeScript SDK provides a powerful toolset for developers to integrate local language models into their applications, enabling seamless integration with various environments and features. Key features: - Use LLMs to respond in chats or predict text completions. - Define functions as tools and turn LLMs into autonomous agents that run locally. - Load, configure, and unload models from memory. - Support for both browser and Node-compatible environments. - Generate embeddings for text. Tech stack: TypeScript Why it matters: LM Studio TypeScript SDK offers a comprehensive solution for developers looking to integrate local language models into their applications, providing a clean and easy-to-use interface for TypeScript/JavaScript developers.
Jupyter Notebook 1,800 stars

onestardao/WFGY

WFGY is heading toward WFGY 5.0 Polaris Protocol, a major open-source release for AI reasoning, RAG, agents, and real-world workflows. Includes Problem Map, Global Debug Card, WFGY 4.0, and the CFV Easter Egg.

### Overview: WFGY is a leading open-source ecosystem focused on AI reasoning, RAG, agents, and real-world workflows. The latest release, WFGY 5.0 Polaris Protocol, is a major upgrade that includes Problem Map 3.0, Global Debug Card, WFGY 4.0, and the CFV Easter Egg. ### Key Features: - **WFGY 5.0 Polaris Protocol**: A comprehensive AI reasoning and verification system. - **Problem Map 3.0**: A fast-track solution for broken AI workflows. - **Global Debug Card**: A tool for debugging complex AI issues. - **WFGY 4.0**: Governance engine for the ecosystem. - **WFGY 3.0**: Frontier reasoning and verification layer. - **WFGY 1.0 to 5.0 Polaris Protocol**: A continuous evolution of AI reasoning tools. ### Tech Stack: - Jupyter Notebook ### Why It Matters: WFGY 5.0 Polaris Protocol represents a significant advancement in AI reasoning and verification, making it a crucial tool for developers and researchers in the field. ### Repository: - Repository: onestardao/WFGY - GitHub Stars: 1800 - GitHub Description: WFGY is heading toward WFGY 5.0 Polaris Protocol, a major open-source release for AI reasoning, RAG, agents, and real-world workflows. Includes Problem Map, Global Debug Card, WFGY 4.0, and the CFV