Project Library

PEX is a Python library designed to create .pex (Python EXecutable) files, which are self-contained, executable Python environments. It simplifies Python application deployment by packaging dependencies and Python distributions into a single portable file, compatible across platforms like Linux and macOS. PEX supports integration with build systems like Pants, Buck, and PyGradle, and allows for creating ephemeral environments, freezing virtualenvs, and building standalone executables for console scripts. It is licensed under Apache2 and can be installed via pip or built using tox. PEX is ideal for developers seeking efficient, portable, and reproducible Python environments.

vLLM is a high-performance, open-source library designed for efficient and scalable large language model (LLM) inference and serving. It features state-of-the-art serving throughput, optimized memory management with **PagedAttention**, and supports advanced techniques like continuous batching, CUDA/HIP graph execution, and various quantization methods (e.g., GPTQ, AWQ, INT4, INT8, FP8). vLLM integrates seamlessly with popular Hugging Face models, offers OpenAI-compatible API servers, and supports distributed inference with tensor and pipeline parallelism. It is highly flexible, supporting a wide range of hardware (NVIDIA, AMD, Intel, TPU, AWS Neuron) and models, including Transformer-based LLMs, Mixture-of-Experts, and multi-modal models. vLLM is community-driven, with contributions from academia and industry, and is backed by sponsors like a16z, Google Cloud, and NVIDIA.

VizTracer is a low-overhead Python tool for tracing, debugging, and profiling code execution, offering detailed function entry/exit timelines with source code. It requires no source code changes for most features and has no package dependencies. VizTracer supports threading, multiprocessing, subprocesses, async, and PyTorch, with a powerful front-end powered by Perfetto that can smoothly render large traces. It works across Linux, macOS, and Windows. Key features include trace filtering, custom event logging, and support for Jupyter notebooks. Installation is via pip, and it can be used via command line, inline scripts, or a VS Code extension. VizTracer is optimized for performance, with minimal overhead, and supports advanced use cases like remote attach and flamegraphs.

Gitstars is a GitHub Stars repository manager designed to help developers efficiently organize and access their starred projects. It addresses the challenge of locating specific repositories among a growing list by offering features like **Your Stars**, which categorizes repositories by Topics and Language for quick retrieval. Additionally, **Gitstars Ranking** highlights the top 100 most-starred projects across various programming languages, updated daily. Other key features include in-app **README.md previews** and direct links to GitHub repositories and project websites. Gitstars simplifies repository management, making it an essential tool for developers to streamline their workflow and discover trending projects.

Checkmate is an open-source, self-hosted monitoring tool designed to track server hardware, uptime, response times, and incidents in real-time. It offers comprehensive monitoring capabilities, including website, page speed, infrastructure, Docker, ping, SSL, and port monitoring, with real-time alerts and reports. The tool integrates with the Capture agent for detailed server insights like CPU, RAM, and disk usage. Built with ReactJS, Node.js, MongoDB, and Docker, Checkmate is optimized for performance, supporting 1000+ monitors with minimal resource usage. It features a user-friendly dashboard, email notifications, and scheduled maintenance. Future updates include global uptime checking, status pages, and enhanced notification options. Checkmate is ideal for businesses and individuals seeking a reliable, customizable monitoring solution.

`aisuite` is a Python library offering a simple, unified interface to interact with multiple Generative AI providers, including Anthropic, AWS, Azure, Google, OpenAI, and more. It provides a standardized API similar to OpenAI’s, enabling developers to seamlessly switch between and compare different large language models (LLMs) without code changes. The library supports chat completions and includes features like tool/function calling, with both manual and automatic execution options. `aisuite` is lightweight, using HTTP endpoints or SDKs for stability, and allows easy integration of new providers through a consistent naming convention. Installation is flexible, supporting base or provider-specific setups. It’s ideal for developers seeking a streamlined way to test and deploy multiple LLMs.

This repository provides comprehensive resources for learning Low Level Design (LLD) and Object-Oriented Design (OOD), tailored for interview preparation. It covers fundamental OOP concepts, SOLID principles, and design patterns (creational, structural, behavioral), along with UML diagrams for system design. The project includes a curated list of LLD interview problems categorized by difficulty (easy, medium, hard), such as designing a parking lot, ATM, or ride-sharing service. Additional resources include recommended books, Coursera courses, and a newsletter for updates. Contributions are encouraged to enhance content and add new problems, making it a collaborative and evolving learning platform.

ArchiSteamFarm (ASF) is a versatile C# application designed for farming Steam trading cards from multiple accounts simultaneously without requiring the Steam client. It supports unlimited accounts, operates on servers or desktop-less machines, and is compatible with Windows, Linux, and macOS via .NET Core. Key features include automatic card farming, offline mode, SteamGuard and 2FA support, trade management, and a plugin system for customization. ASF ensures VAC-free operation, prioritizes security, and offers advanced functionalities like key redemption, gift acceptance, and chat logging. It’s a powerful tool for Steam users seeking efficiency and flexibility in managing their accounts.

The **docker-xiaoya** project offers a streamlined, one-click deployment solution for **Alist** and **Emby** using **Docker Compose**, compatible with Synology, Linux, Windows, and Mac across X86 and Arm architectures. It automates the entire process, including Alist deployment, metadata downloading, and Emby setup, without manual intervention. Key features include automatic updates for Alist cloud data, Emby configurations, and media libraries, support for multiple cloud storage services (Aliyun, Quark, PikPak), and WebDav/TvBox integration. The project ensures a clean system environment by embedding all scripts within Docker images and provides easy customization for advanced users. It’s ideal for creating a seamless media management and streaming experience.

Dependency-Check is a Software Composition Analysis (SCA) tool designed to identify publicly disclosed vulnerabilities in project dependencies by matching them against Common Platform Enumeration (CPE) identifiers and linking to associated CVE entries. It supports multiple integration methods, including CLI, Maven, Gradle, Ant, and Jenkins, and requires Java 11 or higher. The tool relies on the NVD API for vulnerability data, with an API key highly recommended for optimal performance. It supports various technologies like .NET, Go, Ruby, and npm, and can be used in CI environments with caching strategies to manage rate limits. Dependency-Check generates detailed reports and is licensed under Apache 2.0.

Cobalt is a user-friendly, ad-free media downloader designed for simplicity and efficiency. It allows users to download publicly accessible content by pasting a link, ensuring a seamless experience without ads, trackers, or paywalls. The project is structured as a monorepo, containing source code for the API, frontend, and related packages, along with comprehensive documentation for running and protecting instances. Cobalt operates as a proxy, taking zero liability for user actions and never caching content. It emphasizes ethical use, supporting only free and publicly available content, and is not intended for piracy. Contributions are welcome, guided by clear guidelines, and the project is supported by sponsors like RoyaleHosting. Licensed under AGPL-3.0, Cobalt prioritizes transparency and community engagement.

Resume Matcher is a free, open-source AI tool designed to optimize resumes for job applications. It analyzes resumes and job descriptions, extracting relevant keywords and themes to improve ATS (Applicant Tracking System) compatibility. The tool uses advanced machine learning and vector similarity techniques to measure alignment between resumes and job descriptions, providing actionable insights for enhancement. Key features include keyword extraction, readability improvement, and deep resume analysis. Built with Python, it supports local installation, Docker, and Google Colab, with a user-friendly web interface for interactive use. The project encourages community contributions and is licensed under Apache 2.0, making it accessible for developers and job seekers alike.

Olauncher is a minimalist, ad-free Android launcher designed for simplicity and efficiency. It offers a clean, distraction-free interface, focusing on essential functionality while keeping advanced features hidden to maintain its streamlined design. Available for download via F-Droid, the Play Store, or direct APK from GitHub, Olauncher is ideal for users seeking a lightweight and clutter-free home screen experience. Licensed under GNU GPLv3, it emphasizes user privacy and customization without compromising on performance. Explore its features and FAQs through the in-app About page for a deeper understanding of its capabilities.

OpenHaystack is a framework enabling users to create personal Bluetooth tracking tags compatible with Apple's Find My network. It allows tracking of physical objects or Bluetooth-capable devices globally, leveraging nearby iPhones to upload encrypted location data to Apple's servers. The project includes a macOS app for managing and displaying device locations and firmware for deploying tracking beacons on devices like BBC micro:bit or ESP32. OpenHaystack is experimental, requiring macOS 11, and integrates with Apple Mail for location report retrieval. It supports customization and deployment on various Bluetooth-enabled platforms, offering a privacy-focused, decentralized tracking solution.

Moodist is a privacy-focused, open-source web application offering over 75 ambient sounds to enhance focus and relaxation. Key features include persistent sound selection, custom presets, a sleep timer, Pomodoro timer, notepad, and media controls. Users can share sound selections, utilize keyboard shortcuts, and access a simple to-do list (coming soon). Built with TypeScript, React, and Astro, it emphasizes accessibility and clean design. Moodist is completely free, self-hostable, and collects no user data. It supports contributions and is licensed under MIT, with some sounds sourced from third-party providers under different licenses.