Project Library

A development tool for creating powerful AI applications, it provides APIs for plugins and datasets, as well as an interface for quick engineering and visualization operations. For developers and researchers who want to develop applications, Dify provides convenient tools and interfaces to help them build feature-rich AI applications.

10 Lessons to Get Started Building AI Agents

A lightweight web server written in Go. Compared to well-known web servers like Apache and Nginx, its distinctive feature is that it provides a compiled executable file, achieving true out-of-the-box functionality. It offers humanized features such as free HTTPS without any configuration and automatically converting Markdown files into HTML. For building small to medium-sized web services, it is more than sufficient and saves time and effort.

WeChatMsg is a project designed to empower users to take control of their WeChat data, emphasizing the preservation of personal memories and interactions. It supports WeChat 4.0, offering features like local database access, chat interface restoration, and comprehensive data export options (SQLite, HTML, CSV, TXT, Word). The project also includes chat data analysis and visualization, enabling users to generate personalized annual reports. With a focus on user-friendly design, it features a streamlined UI, lower memory usage, and faster export speeds. WeChatMsg is built on the principle of "My Data, My Control," ensuring users can retain and utilize their digital footprints responsibly. It is open-source, encouraging community contributions and fostering innovation in personal AI development.

Swift is a high-performance, modern system programming language designed for safety, speed, and expressiveness. It features a clean syntax, seamless interoperability with C and Objective-C, and memory safety by default. Swift supports a wide range of platforms, including macOS, Ubuntu, Amazon Linux, Debian, Fedora, and Windows, across architectures like x86_64, AArch64, and ARM64. It also enables cross-compilation for targets such as WebAssembly (Wasm). Swift embraces modular design, eliminating headers and code duplication, while offering high-level constructs like objects, protocols, closures, and generics. The project encourages community contributions and provides comprehensive documentation for building, debugging, and contributing to the compiler. Swift aims to foster a diverse and inclusive developer community.

This project is the first semester of the girafe-ai Machine Learning course, designed to provide foundational knowledge in ML. It includes weekly lectures, seminars, and assignments covering key topics such as Naive Bayes, kNN, linear regression, SVM, PCA, trees, ensembles, gradient boosting, and deep learning concepts like backpropagation, dropout, and embeddings. The course features recorded lectures, slides, and practical assignments with deadlines, supplemented by additional materials and literature. Prerequisites and an exam program are provided for structured learning. Authored by Radoslav Neychev and Vladislav Goncharenko, with contributions from several experts, the course is ideal for beginners seeking a comprehensive introduction to machine learning.

The **AI Engineering Hub** is a comprehensive resource for AI practitioners, offering in-depth tutorials on **LLMs and RAGs**, real-world **AI agent** applications, and scalable implementation examples. Designed for all skill levels, it supports learning, experimentation, and project adaptation in AI engineering. Contributors are encouraged to enhance the repository through forks, branches, and pull requests. Stay updated with the latest insights and tutorials by subscribing to the newsletter, which includes a free Data Science eBook. The project is licensed under MIT, fostering open collaboration and community engagement. Connect via issues or direct communication for discussions and suggestions.

The Google Cloud Client Libraries for Go provide Go packages to interact with Google Cloud Platform (GCP) services. These libraries support the two most recent major Go releases and offer seamless authentication using Application Default Credentials (ADC), simplifying integration in GCP environments like Compute Engine, Kubernetes Engine, and App Engine. For local development, authentication can be configured using the Google Cloud CLI or service account key files. The libraries are under active development, with occasional backward-incompatible changes. Contributions are encouraged, following the project’s Contributor Code of Conduct. Comprehensive documentation and quickstart guides are available for various GCP services, including App Engine, Cloud Functions, and Cloud Run.

"Free-for.dev" is a curated list of SaaS, PaaS, IaaS, and other services offering free tiers specifically tailored for developers, system administrators, and DevOps practitioners. It focuses on infrastructure-related tools, excluding self-hosted software and short-term trials. The list emphasizes services with free tiers lasting at least a year and ensures security features like TLS are not restricted to paid plans. Contributions are community-driven, with over 1,600 contributors adding or updating entries via pull requests. Categories include cloud providers, CI/CD, monitoring, APIs, storage, and more, making it a valuable resource for developers seeking cost-effective solutions for their projects.

Maxun is an open-source, no-code web data extraction platform designed to simplify web scraping. Users can train robots in just two minutes to automate data extraction tasks, such as capturing lists, text, or screenshots. It supports pagination, scrolling, and scheduled runs, enabling users to turn websites into APIs or spreadsheets seamlessly. Maxun offers integrations like Google Sheets and plans to add features like layout adaptation and two-factor authentication support. It can be deployed locally via Docker or manually with Node.js, PostgreSQL, MinIO, and Redis. A managed cloud version is also available, handling anti-bot detection, proxy rotation, and CAPTCHA solving for scalable data extraction. Ideal for users seeking efficient, no-code web scraping solutions.

Union is a hyper-efficient zero-knowledge infrastructure layer designed for general message passing, asset transfers, NFTs, and DeFi. It operates without dependencies on trusted third parties, oracles, multi-signatures, or MPC, leveraging Consensus Verification for security. Union is compatible with Cosmos chains via IBC and connects to EVM chains like Ethereum, Berachain, and Arbitrum. Its decentralized governance controls contract upgrades, connections, token configurations, and protocol evolution. Key components include `uniond` (node implementation), `galoisd` (ZK prover), `voyager` (cross-ecosystem relayer), and `hubble` (chain indexer). Built with Go, Rust, and Solidity, Union supports reproducible builds via Nix and offers a TypeScript SDK for interaction. It aims to align priorities among users, validators, and operators through decentralized governance.

LightRAG is a lightweight, efficient Retrieval-Augmented Generation (RAG) system designed for fast and scalable knowledge retrieval and generation. It supports multiple retrieval modes, including local, global, hybrid, and mix, integrating both structured knowledge graphs and unstructured vector search for comprehensive answers. Key features include multi-file type support (PDF, DOC, PPT, CSV), custom knowledge graph integration, and advanced query capabilities with citation functionality. LightRAG is compatible with various storage solutions like Neo4J, PostgreSQL, and Faiss, and supports LLM models from OpenAI, Hugging Face, and Ollama. It also offers a user-friendly GUI for document indexing, querying, and graph visualization. Designed for simplicity and speed, LightRAG is ideal for applications requiring efficient knowledge extraction and generation.

The **LLM Cookbook** is a comprehensive guide for developers aiming to master Large Language Models (LLMs). It adapts Andrew Ng's LLM courses, offering translated, optimized, and practical content tailored for Chinese learners. The project covers essential topics like Prompt Engineering, RAG development, and model fine-tuning, structured into **11 courses** divided into **core** and **elective** categories. Core courses focus on foundational skills, while electives explore advanced techniques like Gradio, LangChain, and RAG systems. The guide includes bilingual code examples, Chinese prompts, and accessible online/PDF formats, making it ideal for Python developers seeking to build LLM-powered applications. Contributions are encouraged to expand the resource further.

CUDA Python provides Pythonic access to NVIDIA’s CUDA platform, enabling GPU programming and high-performance computing. It includes multiple components: `cuda.core` for idiomatic CUDA Runtime access, `cuda.bindings` for low-level CUDA C API bindings, `cuda.cooperative` for device primitives, and `cuda.parallel` for efficient parallel algorithms like sorting and reduction. Additionally, it integrates with Numba for compiling Python code into CUDA kernels. Structured as a metapackage, CUDA Python allows independent installation of subpackages, focusing on developer productivity, ease of use, and reducing the learning curve for CUDA development. It supports end-to-end CUDA development entirely in Python while maintaining compatibility with the latest CUDA features.

FunASR is a comprehensive end-to-end speech recognition toolkit designed to bridge academic research and industrial applications. It supports a wide range of functionalities, including speech recognition (ASR), voice activity detection (VAD), punctuation restoration, language models, speaker verification, and diarization. The toolkit offers pre-trained models like Paraformer-large, known for high accuracy and efficiency, and supports both streaming and non-streaming ASR. FunASR provides easy-to-use scripts, tutorials, and deployment options for inference and fine-tuning, making it accessible for researchers and developers. It also integrates with platforms like ModelScope and Hugging Face, offering a vast collection of models for various languages and tasks.