llama.vim is a powerful Vim plugin offering intelligent code and text completion using a locally running LLM. It supports various LLMs and boasts efficient performance even on low-end hardware thanks to smart context management. Easy to install and configure, the plugin offers customizable completion options, significantly boosting coding productivity.
Read more
Bunster is a project that compiles shell scripts into efficient, static binaries, enhancing portability and security. Instead of simply wrapping scripts, it leverages the Go compiler to translate scripts into native machine code, enabling cross-platform execution (currently Unix only). Bunster primarily supports Bash scripts, with plans to support more shells, and add features like a module system, static asset embedding, password and expiration locks. The project is in its early stages but has already implemented many core features and follows SemVer versioning.
Read more
libmodulor is an opinionated TypeScript library designed to accelerate the creation of high-quality business applications. It employs a layered architecture (UseCase, App, Product, Target), emphasizing type safety and end-to-end testing. Developers are free to use their preferred database, frontend framework, server, and tools; libmodulor provides higher-level primitives to avoid boilerplate code and vendor lock-in. While still under active development, it's already been used in pilot projects.
Read more
edge-tts is a Python library that lets you use Microsoft Edge's online text-to-speech service directly from your Python code or via command-line tools (`edge-tts` and `edge-playback`). No need for Microsoft Edge, Windows, or an API key. It supports multiple languages and voices, offering customization of speed, volume, and pitch. Install via pip and use simple command-line arguments or the Python API.
Read more
Apache Hudi is an open-source data lakehouse platform built on a high-performance open table format for ingesting, indexing, storing, serving, transforming, and managing data across multiple cloud data environments. It supports various data formats and sources, offers atomic commits with rollback/restore, and boasts fast upsert/delete capabilities. A scalable indexing subsystem accelerates queries, while support for snapshot, incremental, and time-travel queries provides flexibility. Integration with metadata stores like Apache Hive Metastore is also included.
Read more
BrowserAI is an open-source project enabling you to run large language models (LLMs) locally within your browser. Prioritizing privacy, all processing happens on your device, eliminating server costs and complex infrastructure. It supports multiple models, including those from MLC and Transformers, leveraging WebGPU for blazing-fast inference. A simple API allows developers to easily integrate text generation, speech recognition, and text-to-speech. Many models are already supported, with a roadmap outlining future enhancements such as advanced RAG capabilities and enterprise features.
Read more
A developer built a retro home computer using just 61 74HCxx logic ICs, 2 6C1008 SRAM chips, and 4 39SF0x0 FLASH chips – boasting 4x the processing power of a Commodore C64! The Minimal 64x4 features 64KB RAM, a 512KB FLASH SSD, VGA output, and a PS/2 port. It has 256 instructions and a complete toolchain including an OS, text editor, assembler, and a Python-like interpreter. The open-source project includes classic games like Tetris and Space Invaders.
Read more
NotepadJS is a minimalist PWA application that mimics the Windows Notepad. The developer aimed to recreate the simplicity and ease of use of the classic Notepad, leveraging the modern File System Access API for local file reading and writing. This project marks the developer's first foray into Vue.js, and contributions in the form of criticism, issues, and pull requests are welcomed.
Read more
Evo is a modern, offline-first version control system designed to simplify development workflows and enhance team collaboration. Say goodbye to complex Git branching and frustrating merge conflicts. Evo uses workspaces instead of branches, supports intelligent merging of JSON and YAML files, and offers built-in authentication and optional commit signing. Its core philosophy is simplicity, efficiency, and reliability. Built with Go, it supports large files and granular permissions.
Read more
Tencent unveils Hunyuan3D 2.0, a cutting-edge large-scale 3D synthesis system generating high-resolution textured 3D assets. Comprising two core components – the Hunyuan3D-DiT shape generation model and the Hunyuan3D-Paint texture synthesis model – it surpasses existing state-of-the-art models in geometry detail, condition alignment, and texture quality. A user-friendly platform, Hunyuan3D-Studio, simplifies manipulation and animation. Inference code and pre-trained models are now open-sourced, accessible via Hugging Face and the official website.
Read more
MoonshotAI has unveiled Kimi k1.5, a new multi-modal large language model trained with reinforcement learning, achieving state-of-the-art results across various benchmarks. Key to Kimi k1.5's success is its 128k context window and improved policy optimization, enabling strong reasoning capabilities without complex techniques like Monte Carlo tree search. It outperforms GPT-4o and Claude Sonnet 3.5 on tests like AIME, MATH-500, and Codeforces, also showing significant improvements in short-context reasoning. Kimi k1.5 will soon be available at https://kimi.ai.
Read more
Ruff is an extremely fast Python linter and code formatter written in Rust. It's 10-100x faster than existing tools like Flake8 and Black, offering drop-in parity with popular tools while boasting built-in caching and automatic fix capabilities. With over 800 built-in rules and support for pyproject.toml, Ruff is used by major open-source projects like FastAPI and Pandas, making it a game-changer for Python development.
Read more
Kronotop is a Redis-compatible, distributed, and transactional document database built on FoundationDB. It supports an MQL-like query language, ACID transactions, and an on-disk storage engine with a primary-standby replication model. While still in early development with an unstable API, it already features ZMap (an ordered key-value store), namespaces, and partial support for Redis data structures (strings and hashes). Future plans include support for more Redis data structures (lists, sets, sorted sets) and a Bucket data structure for storing JSON-like documents with MQL query language and transaction support.
Read more
The AMD ROCm open-source compute platform is seeking community input to determine which GPUs will receive driver support in the future. Currently, ROCm supports select AMD Instinct and Radeon cards, but many users are requesting support for more models, particularly those with 16GB or more VRAM, and reinstatement of support for older AMD GPUs that have lost ROCm compatibility. A GitHub discussion thread has generated significant community engagement, with users actively voting on their desired GPU support.
Read more
WordPecker is a personalized language learning app that combines Duolingo-style lessons with your own curated vocabulary lists. Seamlessly add words from books, articles, or videos, and review them through interactive quizzes and LLM-generated lessons. The app tackles the pain points of vocabulary acquisition in traditional language learning by connecting learning to context, boosting efficiency and retention. Currently featuring multiple question types, future development includes progress tracking, list sharing, and more.
Read more
The DeepSeek team open-sourced its first-generation reasoning models, DeepSeek-R1, and a suite of distilled models. DeepSeek-R1-Zero, trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT), demonstrates remarkable reasoning capabilities, though it has some flaws. DeepSeek-R1 addresses these issues by incorporating cold-start data before RL, achieving performance comparable to OpenAI-o1. Six distilled models based on Llama and Qwen are also open-sourced, with DeepSeek-R1-Distill-Qwen-32B outperforming OpenAI-o1-mini on various benchmarks. The project supports commercial use and provides an online chat website and an OpenAI-compatible API.
Read more
A developer has ported the classic game DOOM into a Microsoft Word document! Using VBA macros and base64 encoding, they embedded the DOOM engine and game data within a Word document, rendering the game in real-time by reading keyboard input. While lacking sound, this unusual project showcases the power of VBA and the developer's creativity, prompting reflection on security and the exploration of software boundaries.
Read more
The open-source project `terraform-provider-hrui` offers a Terraform provider for managing HRUI brand switches (Horaco, Sodola, XikeStor, etc.). It uses goquery to interact with the switch's web UI and has been successfully tested with the Horaco ZX-SWTG124AS. Contributions are welcome, and the project is licensed under the MPL-2.0 license.
Read more
Rhai is an embedded scripting language and evaluation engine for Rust, providing a safe and easy way to add scripting capabilities to any application. It targets all CPU and OS architectures supported by Rust, including WebAssembly, and features a simple, JavaScript-like syntax. Rhai boasts efficient evaluation and tight integration with native Rust functions and types, allowing seamless passing and use of Rust values. Its robust security features protect against malicious attacks such as stack overflows and runaway scripts.
Read more
Many have tried using their Apple device as an access card, but the closed nature of NFC and Wallet ecosystems makes this difficult. However, a Chinese transit card called "China T-Union," officially supported by Apple Wallet, offers a clever workaround. Its unique properties – a non-randomizing UID and unchanging serial number across devices – allow it to be recognized by some UID-based access control systems. While UID authentication is less secure, some systems support it as a fallback. Obtaining the card requires an Alipay account and a biometric travel document. The method is slightly convoluted, but it provides a viable solution for access systems supporting UID authentication.
Read more
This guide provides a comprehensive explanation of Linux network programming, focusing on socket programming. The author notes that many online resources lack clarity and sample codes often only cover the basics, hence the creation of this tutorial, offering clear guidelines and numerous examples. Topics covered include socket types, addressing, APIs (getprotobyname(), getservbyname(), getaddrinfo(), htonl(), htons(), ntohl(), ntohs(), socket(), setsockopt(), bind(), listen(), accept(), connect(), recv(), send(), close()), client-server models (simple HTTP client, TCP-based client-server, multithreaded TCP client-server, UDP-based client-server), advanced techniques (non-blocking sockets, synchronous I/O multiplexing with select() and poll(), broadcasting messages), and secure networking with libcurl and OpenSSL.
Read more
Yek is a fast Rust-based tool designed to read text-based files from a repository or directory, chunk them, and serialize them for Large Language Model (LLM) consumption. It intelligently skips unwanted files (using .gitignore rules and Git history), splits content into chunks based on token count or byte size, and offers extensive configuration options such as custom ignore patterns and file priority rules. Yek aims to streamline the data preparation process for LLMs, boosting efficiency.
Read more
BorrowChecker.jl is a Julia package that simulates Rust-like ownership and borrowing semantics through a macro-based system, performing runtime checks during development and testing to flag memory safety issues. It doesn't guarantee absolute memory safety but aids in writing safer code. The library uses macros like `@own`, `@move`, `@clone` to manage object ownership and `@lifetime` blocks to manage reference lifetimes, while also offering an `@managed` block for automatic borrow checking. While not perfect, it's a valuable tool for developers aiming to improve memory safety in their Julia code.
Read more
llmpeg simplifies using FFmpeg commands via a large language model (LLM). Simply describe your desired video manipulation, and llmpeg generates the corresponding FFmpeg command. For instance, you can easily remove audio from a video. Requires FFmpeg and an OpenAI API key. This project provides a convenient AI-powered tool for video processing.
Read more
Tired of monotonous coding? The `vscode-pets` VS Code extension adds playful pets like cats, dogs, and snakes to your coding environment! Easily installed, it offers various customization options for pets and themes. With community translations, this extension promises a more enjoyable coding experience.
Read more
Dr.TVAM is an inverse rendering framework for tomographic volumetric additive manufacturing (TVAM) built on the Mitsuba renderer. It leverages physically-based differentiable rendering to optimize patterns for TVAM, supporting scattering printing media, arbitrary vial shapes, and projector motions. An improved discretization scheme for the target shape is also included. A command-line tool and comprehensive documentation are provided for easy customization and optimization.
Read more
The c2fj project is a compiler that translates C code into FlipJump, a minimalist instruction set virtual machine. It first compiles C code into a RISC-V ELF file, then translates RISC-V instructions into FlipJump instructions, and finally runs the resulting FlipJump program. This process cleverly leverages FlipJump's space optimization, allowing for efficient compilation even of large C codebases. c2fj supports various compilation flags and provides Makefile support for complex projects. It also includes a minimal C library implementing necessary system calls such as `exit`, `getc`, `putc`, and `sbrk`.
Read more
WestonCB's Nonlinear Optics Sandbox is an interactive platform for studying controlled nonlinear dynamics through the simulation of coupled optical fields. The project uses a specific optical configuration—coupled fundamental/harmonic fields interacting with an adaptive medium—as an idealized model to explore how optimization processes can guide complex nonlinear systems towards desired states. It combines real-time visualization, GPU-accelerated physics simulation, and gradient-based optimization, offering a novel tool for researching nonlinear dynamics and the control of complex systems.
Read more
Nyxelf is a powerful tool designed for analyzing malicious Linux ELF binaries. It combines static analysis techniques using tools like readelf, objdump, and pyelftools with dynamic analysis within a secure QEMU-based sandbox. Features include UPX unpacking, syscall tracing, process/file activity monitoring, and an intuitive GUI powered by pywebview. JSON output supports automated workflows, making Nyxelf ideal for security researchers and reverse engineers.
Read more
A programmer spent over a year solving the first 100 Project Euler problems using 100 different programming languages! From common languages like Python and Java to obscure esoteric languages, the sheer dedication and programming prowess is impressive. This project showcases deep understanding of various programming paradigms and offers a valuable learning experience for programmers of all levels.
Read more