Projects

### Overview: Agentic Awesome Skills is a comprehensive library of 2,705+ installable `SKILL.md` playbooks designed to help agents discover, select, validate, and plan stacks. The core includes a CLI, local MCP, catalog, plugins, and Workbench, backed by 2,600+ agentic skills. ### Key features: - **CLI**: Command-line interface for quick access. - **Local MCP**: Agent-owned selection capabilities. - **Stack Validation**: Ensures stack compatibility and correctness. - **Plan Preview**: Preview and inspect a plan before making changes. ### Tech stack: Python ### Why it matters: Agentic Awesome Skills provides a structured way for agents to find, select, and validate skills, ensuring a consistent and reliable skill set across projects. The library is designed to be installable and compatible with various tools, making it a valuable resource for developers and agents alike. ### Repository: sickn33/agentic-awesome-skills ### GitHub stars: 47,779

### Overview: CowAgent is an open-source personal AI assistant designed to proactively manage tasks, control your computer, and enhance your knowledge base. It is built with a lightweight, extensible architecture, supporting multiple models and platforms, and continuously evolves through self-evolution. ### Key Features: - **Task Management**: Plans and executes tasks efficiently. - **Skill Creation**: Runs Skills to automate tasks and improve efficiency. - **Knowledge Base**: Builds a personal knowledge base and long-term memory. - **Multi-Agent Team**: Forms a team of Agents with shared conversations and collaborative skills. - **Self-Evolution**: Continuously improves through self-review and learning from interactions. ### Tech Stack: - Python ### Why It Matters: CowAgent is a powerful and flexible AI assistant that can help you manage your tasks, enhance your knowledge, and collaborate with other AI agents. Its lightweight and extensible design makes it easy to integrate with various models and platforms, making it a valuable tool for personal productivity and knowledge management. Repository: zhayujie/CowAgent Primary Language: Python GitHub Stars: 47,446

### Overview: Apache Airflow is a platform designed to programmatically author, schedule, and monitor workflows. It provides a robust environment for automating tasks and processes, making it an essential tool for developers and IT professionals alike. ### Key Features: - **Workflow Management**: Author, schedule, and monitor workflows. - **Task Execution**: Execute tasks on various platforms. - **API Interface**: Provides a RESTful API for easy integration with other systems. - **Security**: Supports secure authentication and authorization. - **Scalability**: Designed to handle large-scale deployments. ### Tech Stack: Python ### Why It Matters: Apache Airflow is crucial for anyone looking to automate their workflows and processes. Its ability to author, schedule, and monitor tasks makes it invaluable for developers, IT professionals, and anyone involved in system automation. With its robust security features and scalability, Airflow is a must-have tool for any project requiring automated workflows.

Overview: Paperless-ngx is a document management system that converts physical documents into searchable online archives, aiming to reduce paper usage. Key features: - Scan and index physical documents - Archive and store documents securely - Searchable online archives for easy access Tech stack: Node.js, React, TypeScript, Docker, and other modern web technologies. Why it matters: Simplifies document management, reduces paper waste, and improves accessibility to important documents. Repository: paperless-ngx/paperless-ngx GitHub stars: 46,816 GitHub description: A community-supported supercharged document management system: scan, index and archive all your documents.

### Overview: Milvus is a high-performance, cloud-native vector database designed for scalable vector search. Built for AI applications, it efficiently organizes and searches vast amounts of unstructured data, such as text, images, and multi-modal information. ### Key features: - **Scalable Vector Database**: Milvus is optimized for handling billions of vectors across tens of thousands of search queries. - **Hardware Acceleration**: Implemented with hardware acceleration for CPU/GPU to achieve best-in-class performance. - **Fully Distributed Architecture**: Supports horizontal scaling and real-time streaming updates. - **Standalone Mode**: Available for single-machine deployment, ideal for quickstart in Python with `pip install`. ### Tech stack: Go and C++ Milvus is written in Go and C++ for efficient performance and hardware acceleration. ### Why it matters: Milvus is crucial for applications requiring scalable and high-performance vector search, enabling AI applications to efficiently organize and search vast amounts of unstructured data. Its fully-distributed and K8s-native architecture makes it highly scalable and capable of handling real-time updates and large-scale deployments. ### Repository: milvus-io/milvus - **GitHub Stars**: 46,408 - **Primary Language**: Go - **GitHub Description**: Milvus is a high-performance vector database built for scalable vector ANN search. ### Quickstart: To get started with Milvus, install the Python SDK using `pip install -U pymilvus`. This allows you to create a client for interacting with the Milvus database

### Overview: Coqui.ai/TTS is a powerful library for advanced Text-to-Speech generation, featuring 16 languages and improved performance. It supports fine-tuning across various languages and includes utilities for dataset analysis and inference. ### Key features: - 📝 16 languages support - 🚀 Pretrained models in over 1100 languages - 🚀 Fine-tuning code example recipes - 🚀 Stream inference with low latency - 🚀 Supports unconstrained voice cloning - 🚀 Fairseq models compatibility - 🚀 Tortoise inference speed enhancement ### Tech stack: Python ### Why it matters: Coqui.ai/TTS is a crucial tool for developers and researchers looking to create high-quality text-to-speech applications, expanding the capabilities of their projects with advanced language support and improved performance.

Overview: Streamlit is a Python library that simplifies the process of building and deploying data applications. It allows developers to create interactive web apps quickly and easily, without the need for extensive coding. Key features: - Streamlit lets you write Python scripts that can be converted into interactive web applications. - It supports live editing, allowing you to see changes in real-time. - It includes a wide range of widgets for creating dynamic user interfaces. - It provides a community cloud platform for deploying and managing apps. Tech stack: Python Why it matters: Streamlit makes it simple and efficient to build data applications, enabling developers to focus on the core functionality of their projects rather than the complexities of web development. Its fast prototyping capabilities and open-source nature make it a popular choice for rapid development and collaboration.

### Overview: Jan is an open-source alternative to ChatGPT that runs 100% offline on your computer, providing full control and privacy. ### Key features: - **Local AI Models**: Download and run LLMs (Llama, Gemma, Qwen, GPT-oss etc.) from HuggingFace. - **Cloud Integration**: Connect to GPT models via OpenAI, Claude models via Anthropic, Mistral, Groq, MiniMax, and others. - **Custom Assistants**: Create specialized AI assistants for your tasks. - **OpenAI-Compatible API**: Local server at `localhost:1337` for other applications. - **Model Context Protocol**: MCP integration for agentic capabilities. - **Privacy First**: Everything runs locally when you want it to. ### Tech stack: Rust ### Why it matters: Jan offers an open-source alternative to ChatGPT that provides full control and privacy, making it a valuable tool for developers and AI enthusiasts.

### Overview: Herdr is the runtime environment where your coding agents live. It seamlessly integrates with your existing development tools, ensuring that your work continues uninterrupted. Whether you're working on multiple machines or need to switch between them, Herdr keeps your sessions intact and allows you to focus on your code without interruption. ### Key Features: - **Background Agent Persistence**: Herdr keeps your terminals running in the background, ensuring that your work continues even when you close the client or lose your SSH connection. After a server or machine restart, Herdr restores the saved layout, allowing you to resume your sessions seamlessly. - **Unified Workspace**: Herdr allows you to manage multiple machines in a single window, combining local work with remote machines. This feature ensures that your local work and remote environments are always accessible. - **Agent Awareness**: Every pane is marked as working, blocked, or idle, ensuring that you know exactly what is happening in your development environment. - **Agent Native Integration**: Agents drive Herdr through the CLI and socket API, allowing you to spawn panes, prompt each other, and wait for answers. This ensures that your agents are always up-to-date and performant. - **Agent Support**: Herdr supports a wide range of agents, including Claude, Codex, Cursor, Opencode, and more. It doesn’t wrap or replace these agents, allowing you to use them as intended. - **First-Class Support for Agents**: Herdr supports plugins, allowing you to extend the functionality of your panes and workflows. This makes Herdr highly customizable and adaptable to your specific

Overview: Sub2API is an open-source gateway service for AI APIs, allowing seamless integration of subscriptions from providers like Claude, OpenAI, and Gemini. It supports carpooling to efficiently share costs and allows for the use of native tools without any modifications. Key features: - Seamless integration of subscriptions from multiple AI providers - Support for carpooling to share costs - Native tool compatibility without modifications Tech stack: Go, Vue, PostgreSQL, Redis, Docker Why it matters: Sub2API simplifies the management and cost-efficiency of AI API subscriptions, enabling developers to focus on innovation rather than infrastructure.

### Overview: Kong Gateway is a cloud-native, platform-agnostic, scalable API and AI Gateway that distinguishes itself with its high performance and extensibility via plugins. It provides advanced AI traffic capabilities with multi-LLM support, semantic security, MCP traffic security, and analytics, making it a central layer for orchestrating microservices or conventional API traffic. ### Key Features: - **Advanced Routing & Load Balancing**: Configurable and scalable routing and load balancing. - **Multi-LLM Support**: Advanced AI capabilities with support for multiple language models. - **Semantic Security**: Enhanced security features based on language understanding. - **MCP Traffic Security & Analytics**: Advanced security and analytics for MCP traffic. - **Authentication & Authorization**: Comprehensive authentication and authorization features. - **Health Checkers**: Automated health checks for services. - **API Management**: Centralized management of APIs, including routing, load balancing, and security. ### Tech Stack: Lua ### Why it Matters: Kong Gateway serves as the central layer for orchestrating microservices or conventional API traffic, providing enhanced security, AI capabilities, and advanced routing features. Its Lua-based architecture ensures high performance and extensibility, making it a valuable tool for modern API management and AI gateway solutions.

Overview: Ray is an AI compute engine that provides a unified framework for scaling Python and AI applications from a laptop to a cluster. It consists of a core distributed runtime and a set of AI libraries for simplifying ML compute. Key features: - Ray Core: Tasks, Actors, Objects - AI Libraries: Data, Train, Tune, RLlib, Serve Tech stack: Python Why it matters: Ray simplifies the process of scaling AI applications by providing a unified framework that can be used on any machine, cluster, cloud provider, or Kubernetes, and features a growing ecosystem of community integrations.

### Overview: Gradio is a Python package that simplifies the process of building machine learning web applications. It allows users to create demos or web applications for their models, APIs, or Python functions in a few lines of code without needing JavaScript, CSS, or web hosting experience. ### Key features: - **Easy Setup**: No need for JavaScript, CSS, or web hosting. - **Built-in Sharing**: Share your demo or web application with a single line of code. - **Pythonic Interface**: Use Python's built-in functions and libraries. - **Cross-Platform**: Works on Windows, macOS, and Linux. - **Extensible**: Can be extended with custom components and integrations. ### Tech stack: Python ### Why it matters: Gradio makes it easy to create interactive machine learning demos and web applications without the need for complex setup or web development skills. It simplifies the process of building and sharing models, APIs, or custom functions into user-friendly web interfaces, making it ideal for developers and data scientists looking to showcase their work.

### Overview: This tutorial provides a comprehensive set of examples and explanations for beginners to TensorFlow, covering both TensorFlow 1.x and 2.x versions. The repository includes Jupyter Notebooks and source codes for various TensorFlow operations, including basic models, neural networks, and more. The examples are designed to be easily understandable and include both raw implementations and modern TensorFlow API practices. ### Key Features: - **TensorFlow 1.x & 2.x Examples**: Includes both TensorFlow 1.x and 2.x versions, providing a seamless learning experience. - **Practical Examples**: Covers topics such as linear regression, logistic regression, word2vec, gradient boosted decision trees, and neural networks. - **Source Code and Explanation**: Provides both Jupyter Notebooks and source codes with detailed explanations for better understanding. ### Tech Stack: - TensorFlow 1.x & 2.x ### Why It Matters: - Ideal for beginners looking to understand TensorFlow through practical examples. - Covers a wide range of TensorFlow operations, from basic models to advanced neural networks. - Provides a clear and concise learning path suitable for both beginners and intermediate learners. ### Repository: - Repository: aymericdamien/TensorFlow-Examples - Stars: 43,740 - Description: TensorFlow Tutorial and Examples for Beginners (support TF v1 & v2) ### Tutorial Index: - **Prerequisite**: Introduces the basics of machine learning and the MNIST dataset. - **Introduction**: Includes examples for beginners to understand TensorFlow. - **Basic Models**: Covers linear regression, logistic regression, word2vec

Overview: LangGraph is a low-level orchestration framework designed for building resilient and stateful agents, enabling long-running workflows with durable execution, human-in-the-loop capabilities, comprehensive memory management, and advanced debugging tools. Key features: - **Durable Execution**: Agents persist through failures and resume from the last checkpoint. - **Human-in-the-Loop**: Seamlessly incorporate human oversight for modifications during execution. - **Comprehensive Memory**: Create stateful agents with both short-term working memory and long-term persistent memory. - **Debugging with LangSmith**: Visualize agent behavior with detailed runtime metrics and state transitions. - **Production-Ready Deployment**: Deploy complex workflows with scalable infrastructure. Tech stack: Python Why it matters: LangGraph provides essential infrastructure for building long-running, stateful agents, making it a crucial tool for developers working on AI applications and language models.

### Overview: DeepSpeed is a comprehensive deep learning optimization library designed to simplify, enhance, and accelerate distributed training and inference. It is built on Python and leverages the Apache License 2.0 for its open-source nature, making it accessible to a wide range of developers. ### Key Features: - **Distributed Training**: Automatically manages and optimizes the training process across multiple GPUs or nodes. - **Inference Efficiency**: Simplifies the inference process, making it faster and more efficient. - **Model Parallelism**: Facilitates model parallelism for better performance. - **Low-Precision Training**: Supports low-precision training for faster computation. - **Model Compilation**: Compiles models for optimal performance on specific hardware. ### Tech Stack: - Python ### Why It Matters: DeepSpeed is crucial for developers looking to optimize their deep learning models for large-scale distributed training and inference. It simplifies complex tasks, reduces development time, and ensures efficient use of computational resources, making it an indispensable tool in the modern deep learning landscape. ### Repository: deepspeedai/DeepSpeed - **Primary Language**: Python - **GitHub Stars**: 43,239 - **GitHub Description**: DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective. - **GitHub Description (English)**: DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective. - **Latest News**: - [2026/05] Using Muon Optimizer with DeepSpeed

### Overview: Agno is a comprehensive framework and runtime for agent platforms, allowing developers to build, run, and manage their agent stacks with ease. ### Key Features: - **Build Agent Platform**: Use the Agno SDK to create and deploy agent platforms. - **Run Agent Platform**: Utilize the AgentOS runtime to execute agent applications. - **Manage Platform**: Access the AgentOS UI for comprehensive management and monitoring. ### Tech Stack: Python ### Why It Matters: Agno empowers developers to build, run, and manage agent platforms with ease, ensuring control over data, memory, and security. It integrates seamlessly with various tools and platforms, making it a versatile solution for agent development. ### Repository: agno-agi/agno ### GitHub Stars: 42,781 ### Primary Language: Python ### GitHub Description: Build, run, and manage agent platforms.

### Overview: AI Learning: 数据分析+机器学习实战+线性代数+PyTorch+NLTK+TF2 ### Key features: - 实战项目 - 线性代数 - PyTorch - NLTK - TensorFlow 2.x - 数据分析 - 机器学习实战 ### Tech stack: Python ### Why it matters: 通过本项目，你可以系统地学习数据分析、机器学习、线性代数、PyTorch、NLTK和TensorFlow 2.x等技术，掌握实战项目开发技能，提升编程能力。

Overview: 微舆：一个从0实现的多智能体舆情分析系统，旨在打破信息茧房，还原舆情原貌，预测未来走向，辅助决策。用户只需像聊天一样提出分析需求，智能体开始全自动分析国内外30+主流社媒与数百万条大众评论。 Key features: - 实时监控国内外关键社媒 - 融合微调模型、统计模型等多模型协同工作，确保深度、准度与多维视角 - 多模态解析短视频内容，提取结构化多模态信息 - 强大多模态能力，全面掌握舆情动态 - 强化Agent协作机制，避免同质化思维，催生高质量集体智能与决策支持 - 公私域数据无缝融合，提供外部趋势+内部洞察的分析能力 - 轻量化与高扩展性框架，基于纯Python模块化设计，实现一键式部署 Tech stack: Python Why it matters: 微舆通过融合多种智能体和多模态分析能力，提供了一个简洁通用的数据分析引擎，能够有效辅助决策和预测未来走向。相比同类产品，微舆拥有六大优势：AI驱动的全域监控、超越LLM的复合分析引擎、强大的多模态能力、Agent“论坛”协作机制、公私域数据无缝融合以及轻量化与高扩展性框架。这使得微舆成为驱动一切业务场景的简洁通用的数据分析引擎。

Overview: The anthropics/claude-plugins-official repository is a curated directory of high-quality plugins for Claude Code, managed by Anthropic. It includes both internal and external plugins, each following a standardized structure. Key features: - Curated collection of high-quality plugins - Support for both internal and external plugins - Standardized plugin structure for consistency Tech stack: Python (primary language, though plugins can be in various languages) Why it matters: It provides a trusted source for plugins that enhance the functionality of Claude Code, ensuring quality and security for users.