AI Trending

Project directory

Projects

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

Python 894 stars

pocketpaw/pocketpaw

Your AI agent in 30 seconds. Not 30 hours. Self-hosted, open-source personal AI with desktop installer, multi-agent Command Center(Deep Work), and 7-layer security. Anthropic, OpenAI, or Ollama.

Overview: PocketPaw is an AI agent that runs on your machine, providing quick access to AI capabilities without relying on cloud services. Key features: - Self-hosted, open-source personal AI - Desktop installer for Windows, macOS, and Linux - Multi-agent Command Center for advanced management - 7-layer security - Talks to you over Discord, Slack, WhatsApp, Telegram, or the browser Tech stack: Python Why it matters: PocketPaw offers a self-hosted solution for AI agents, ensuring data stays on your machine and eliminating the need for cloud services, making it ideal for users who prefer to control their data and privacy. Repository: pocketpaw/pocketpaw GitHub stars: 894 GitHub description: Your AI agent in 30 seconds. Not 30 hours. Self-hosted, open-source personal AI with desktop installer, multi-agent Command Center(Deep Work), and 7-layer security. Anthropic, OpenAI, or Ollama.
TypeScript 894 stars

vostride/agent-qa

Open-source self-improving QA agent for software teams. A test harness with memory. Write tests in natural language for web and mobile. agent-qa learns from every run, adapts to UI changes, and catches regressions before you ship.

### Overview: The **agent-qa** is an open-source self-improving QA agent designed for software teams. It uses natural language tests to write and run tests for web and mobile applications. The agent learns from every run, adapts to UI changes, and identifies regressions before deployment. It integrates with your existing development tools and leverages the **mcp** and **skills** to enhance its effectiveness. ### Key Features: - **Self-Improvement:** Validates lessons from runs, retires stale knowledge, and uses feedback to correct mistakes. - **CLI Integration:** Allows testing through the command line. - **Dashboard Access:** Provides a visual interface for monitoring and managing tests. - **Memory Integration:** Keeps track of facts and working procedures as reviewable files in your repository. ### Tech Stack: - TypeScript ### Why It Matters: The **agent-qa** is a powerful tool for ensuring the quality of software releases. By continuously learning from each test run, it helps teams catch regressions and improve their testing practices. This makes the development process more efficient and reliable, ultimately leading to better software products. ### Repository: Repository: **vostride/agent-qa** GitHub Stars: **891** GitHub Description: Open-source self-improving QA agent for software teams. A test harness with memory. Write tests in natural language for web and mobile. agent-qa learns from every run, adapts to UI changes, and catches regressions before you ship. ### Primary Language: TypeScript ### GitHub Description: Open-source self-improving QA agent for
Kotlin 894 stars

wangzhishou/OneBox

A free AI agent & toolbox for Android | 一站式安卓 AI 智能体 & 工具箱

### Overview: OneBox is a free AI agent and toolbox for Android, providing a comprehensive suite of tools and features to streamline your daily tasks. With a focus on efficiency and user-friendliness, OneBox integrates seamlessly into your Android ecosystem, offering a wide range of functionalities without the need for additional installations. ### Key Features: - **90+ In-App Tools**: Drive productivity with a vast array of local tools for image processing, document management, and more. - **Built-in AI Assistant**: Simplify tasks by just stating what you need, with the AI agent handling the rest. - **Local Execution**: All tools run on-device, ensuring privacy and security. - **Picture-in-Picture Functionality**: Keep your agent session active in the background, allowing you to monitor progress without leaving the app. - **Multiple Models**: Supports multiple LLMs, including custom AI engines and prompt management. - **Bring Your Own Key**: Use any OpenAI-compatible API key for free, with full access to the assistant and agent. - **Local Models**: Download and use local models for offline chat, completely free of network, account, or token costs. - **Transparent by Source**: Every line of code is publicly available, ensuring transparency and verifiability. ### Tech Stack: Kotlin ### Why It Matters: OneBox simplifies complex tasks by providing a single, intuitive interface that integrates seamlessly with your existing Android ecosystem. Its AI-driven tools and local execution capabilities make it an indispensable tool for anyone looking to boost productivity and efficiency on their Android device.
Unknown 893 stars

ZhongqunZHANG/awesome-6d-object

Awesome work on object 6 DoF pose estimation

### Overview: This repository contains curated resources for 6D object pose estimation and 3D object reconstruction from a single view. It includes evaluation papers, journal and conference papers, and open-source demos. ### Key features: - arXiv Papers - Journal Papers - TPAMI / IJCV - Others - Conference Papers - 2024: CVPR - 2023: CVPR, ICCV, IROS, ICLR - 2022: CVPR, ECCV, ICLR, IROS, Others - 2021: CVPR, ICCV, IROS, Others - 2020: CVPR, ECCV, Others - 2019: CVPR, ICCV, Others - 2018: CVPR, ECCV, Others - 2017: CVPR, ICCV, Others - 2016: CVPR, Others - 2015: CVPR, ICCV, Others - 2014: CVPR, Others ### Tech stack: C++ and Python ### Why it matters: This repository is a comprehensive resource for researchers and developers interested in 6D object pose estimation and 3D reconstruction. It provides up-to-date papers, demos, and datasets, making it a valuable tool for anyone working in this field.
Python 893 stars

alirezarezvani/claude-code-skill-factory

Claude Code Skill Factory — A powerful open-source toolkit for building and deploying production-ready Claude Skills, Code Agents, custom Slash Commands, and LLM Prompts at scale. Easily generate structured skill templates, automate workflow integration, and accelerate AI agent development with a clean, developer-friendly setup.

### Overview: Claude Code Skill Factory is a comprehensive toolkit designed to help developers quickly and efficiently create production-ready Claude Skills, Code Agents, custom Slash Commands, and LLM Prompts at scale. This open-source project offers templates, examples, and powerful prompt engineering systems, making it easier to build and deploy AI agents across various platforms. ### Key Features: - **Skills Factory**: Generates complete, production-ready Claude Skills with proper frontmatter, Python implementations, sample input/output, documentation, and ZIP packages. - **Agents Factory**: Creates specialized Claude Code Agents with enhanced YAML frontmatter, MCP integration, auto-invocation, and tool access configuration. - **Prompt Factory**: Generates mega-prompts for any role with 69 professional presets, multiple output formats, quality validation, and core & advanced modes. - **Slash Commands**: Provides interactive builders for building skills, agents, and prompts, and ready-made skills for quick integration. - **Interactive Agents**: Offers specialized agents for orchestrating the creation of skills, prompts, and hooks, and a guide agent for skill-building. ### Tech Stack: Python ### Why It Matters: Claude Code Skill Factory simplifies the process of creating and deploying AI agents, making it easier for developers to integrate Claude AI into their projects and streamline workflows. The toolkit's clean, developer-friendly setup and extensive documentation make it an invaluable resource for anyone looking to accelerate AI agent development.
Python 893 stars

harleyszhang/llm_note

LLM notes, including model inference, transformer model structure, and llm framework code analysis notes.

### Overview: This repository contains notes and insights on large language models (LLMs), including model inference, transformer model structure, and framework analysis for LLMs. The primary language used is Python, and the project has garnered 893 GitHub stars. The README excerpt highlights the creation of a lightweight inference framework using Triton and PyTorch, focusing on practical applications and optimization techniques. ### Key features: - **Model Inference**: Notes on how to perform inference on large language models. - **Transformer Model Structure**: Detailed explanations and implementation of transformer models, including llama1, llama3, and gpt1-3. - **LLM Quantization and Compression**: Insights into quantization techniques and algorithms for improving model performance. - **LLM Optimization**: Techniques for optimizing LLM inference, including algorithms, parallel acceleration, and performance analysis. - **High Performance Computing**: Notes on using Triton for high-performance computing, including kernel development and optimization. ### Tech Stack: Python ### Why it Matters: This project provides practical, hands-on knowledge on how to build and optimize large language models, making it valuable for developers looking to improve their skills in this field.
Go 893 stars

kubernetes-sigs/lws

LeaderWorkerSet & DisaggregatedSet: Kubernetes APIs for Distributed and Disaggregated AI Serving

**Overview:** LeaderWorkerSet (LWS) is an API designed for deploying a group of pods as a unit of replication, addressing common deployment patterns of AI/ML inference workloads, particularly multi-host inference workloads where the LLM will be sharded and run across multiple devices on multiple nodes. It aims to optimize real-world serving frameworks and disaggregated architectures, achieving state-of-the-art performance across hardware accelerators. **Key features:** - **LeaderWorkerSet (LWS):** Supports a tightly managed group of pods that represent a "super pod." Each pod in the group has a unique index, and pods can be created in parallel. The group can be scaled, rolled out, and updated as a unit, with all pods in the group being updated together. It supports topology-aware placement and all-or-nothing restart for failure handling. - **DisaggregatedSet (DS):** Built on LWS, DS supports advanced disaggregated workload deployment with support for autoscaling, rollouts, and failure handling. It optimizes for workloads where different phases run on separate infrastructure. **Tech stack:** Go **Why it matters:** LWS and DS are critical for optimizing AI/ML inference workloads, especially in multi-host environments, ensuring efficient and scalable deployment across multiple nodes and hardware accelerators. **Repository:** kubernetes-sigs/lws **GitHub stars:** 882 **GitHub description:** The LeaderWorkerSet and DisaggregatedSet APIs are designed to address common deployment patterns of AI/ML inference workloads, particularly multi-host inference workloads
Python 892 stars

promptise-com/Foundry

The foundation layer for agentic intelligence.

Overview: Foundry is the foundational layer for agentic intelligence, providing agents that think, safe ways for them to act, and a system to keep them running and accountable. Key features: - Agents that think - Safe access to tools and data - Continuous operation and accountability Tech stack: Python Why it matters: Foundry simplifies the process of building agentic systems, allowing developers to focus on the core functionality without the need for extensive integration or rewriting.
Python 891 stars

dondai1234/master-fetch

MCP server for web fetching with Cloudflare bypass, Trafilatura extraction, and smart routing. Free, self-hosted, no API keys.

### Overview: Hound is a powerful, self-hosted MCP server designed for web fetching and bypassing Cloudflare's anti-bot measures. It offers a free, no-key, and self-hosted alternative to more expensive alternatives, providing a single MCP server and warm browser for all your web scraping needs. ### Key Features: - **Free, Self-Hosted, No API Keys**: Hound is completely free to use and can be hosted on any machine, eliminating the need for API keys. - **Multi-Engine Support**: Hound uses multiple engines for better accuracy and recall, including Serper, Tavily, Exa, Firecrawl, and TinyFish. - **Intent-Aware Multi-Query Fan-Out**: Hound detects query intent and generates expanded query variants, improving recall and authority signals. - **Six-Signal Ranking**: Cross-variant consensus, domain reputation, answer-signal scoring, title relevance, URL relevance, and result diversity. - **Stealth Engine**: Hound's next-gen stealth engine bypasses Cloudflare Turnstile, CF Turnstile on CanadianInsider, Medium, StackOverflow, NowSecure, Glassdoor, and DataDome. - **No Scrapling**: All scrapling functionality is replaced with Hound's own modules, making it more efficient and secure. ### Tech Stack: - Python ### Why It Matters: Hound is a reliable, self-hosted solution for web scraping and bypassing Cloudflare's anti-bot measures, offering a free, no-key alternative to more expensive alternatives. Its multi-engine support, intent-aware multi-query
Python 891 stars

nolabs-ai/deepfabric

Generate High-Quality Synthetics, Train, Measure, and Evaluate in a Single Pipeline

### Overview: DeepFabric is a comprehensive pipeline for generating high-quality synthetic training data for language models and agent evaluations. By combining reasoning traces with tool-calling patterns, it creates domain-specific datasets that teach models to think, plan, and act effectively, call tools correctly, and conform to strict schema structures. ### Key Features: - **Topic Graph Generation**: Unique topic graph algorithms ensure high diversity yet domain-anchored relevance. - **Constraint Handling**: Ensures generated samples adhere to structured schema, variable constraints, and execution correctness. - **Tool Mocking**: Directly imports from MCP server schemas and automatically mocks real interfaces. - **Automatic Upload**: Dataset can be uploaded to Hugging Face and imported into popular training frameworks like TRL, Unsloth, and Axolotl. - **Evaluation Engine**: Built-in evaluation engine assesses model performance on unseen tasks derived from training splits. ### Tech Stack: - Python ### Why It Matters: DeepFabric stands out by ensuring high diversity yet domain-anchored relevance through unique topic graph generation algorithms. This guides sample creation to cover all necessary subtopics while avoiding redundancy, which is where other tools often fall short, resulting in model overfit.
TypeScript 890 stars

PragmaticMachineLearning/probly

### Overview: Probly is a next-generation spreadsheet application that integrates traditional spreadsheet capabilities with AI-driven analysis, Python computation, and document processing. It offers smart spreadsheets, Python integration, data visualization, AI-powered insights, and document processing features. ### Key Features: - **Smart Spreadsheet**: Full-featured spreadsheet with formula support and intelligent cell suggestions. - **Python Integration**: Run Python code directly in your browser using WebAssembly. - **Data Visualization**: Create interactive charts and visualizations. - **AI-Powered Analysis**: Get intelligent insights and automated trend analysis. - **Document Processing**: Extract and process data from images. - **Prompt Library**: Access predefined analysis templates and save custom prompts. ### Tech Stack: Next.js 14 with TypeScript and React for the frontend, Pyodide (Python compiled to WebAssembly) for local execution, and OpenAI API for AI integration. ### Why It Matters: Probly simplifies data analysis by combining the power of traditional spreadsheets with advanced AI and Python capabilities, making it easier and more efficient for users to extract insights and perform complex analyses.
Python 890 stars

TickDB/tickdb-unified-realtime-marketdata-api

TickDB: AI-native real time stock API and market data API for US/HK/CN stocks, forex, futures, crypto, indices and commodities. Skill, CLI, MCP, REST API and WebSocket. AI 原生实时股票 API 与金融行情数据 API，覆盖美股、港股、A股、外汇、期货、指数、大宗商品、加密货币。

### Overview: TickDB 是一个面向开发者、AI 代理和多市场金融应用的 AI 原生实时行情数据 API。 ### Key features: - 统一接入：一套 API 覆盖外汇、贵金属、指数、美股、港股、A 股、中国及香港期货、加密货币等多个金融市场的实时与历史行情数据。 - 实时数据：通过 WebSocket 的流式推送，端到端延迟约 10-50ms。 - 公司财务与基本面：提供美股、港股、A 股财报、PE 估值、营收构成、股东与公司事件查询。 - AI 原生：官方提供 Skill / MCP / CLI 三档 AI 接入，AI Agent 与编码助手开箱即用。 - 开发者友好：RESTful API + WebSocket，结构化 JSON 响应，完整文档与多语言示例。 - 全球覆盖：覆盖股票、外汇、贵金属、指数、中国及香港期货与加密货币；品种及数量以产品目录返回为准。 - 免费开始：无需信用卡，立即获取 API 密钥。 ### Tech stack: Python ### Why it matters: TickDB 为开发者、AI 代理和多市场金融应用提供了一个统一的实时行情数据接入点，简化了数据管理与集成过程，提高了开发效率和业务灵活性。
Python 890 stars

explosion/curated-transformers

🤖 A PyTorch library of curated Transformer models and their composable components

Overview: Curated Transformers is a PyTorch library that provides state-of-the-art transformer models composed from reusable components, making it easy to implement new architectures and integrate with existing type-checked code. Key features: - Supports state-of-the-art transformer models, including LLMs such as Falcon, Llama, and Dolly v2. - Composes models from a set of reusable building blocks, providing many benefits: implementing a feature or bugfix benefits all models, supports 4/8-bit inference, and can be integrated with PyTorch's `meta` device. - Consistent type annotations of all public APIs, providing great coding support and easy integration with existing type-checked code. - Easy to study the building blocks, making it suitable for education. - Minimal dependencies. Tech stack: Python Why it matters: Curated Transformers is a powerful and versatile library for developers looking to build and integrate transformer models efficiently and effectively.
Python 890 stars

luispedro/mahotas

Computer Vision in Python

Overview: Mahotas is a Python library for computer vision that includes fast algorithms implemented in C++ for efficiency. It supports various image processing and computer vision tasks. Key features: - Fast algorithms implemented in C++ - Over 100 functions for image processing and computer vision - Supports Python 3.11 through 3.15 - Notable algorithms include watershed, convex points calculations, hit & miss, thinning, Zernike & Haralick, LBP, SURF, thresholding, convolution, Sobel edge detection, spline interpolation, SLIC superpixels Tech stack: Python, C++ Why it matters: Mahotas is essential for developers looking to implement efficient computer vision algorithms in Python, making it a valuable tool for image processing and computer vision tasks.
TypeScript 889 stars

rodrigopivi/Chatito

🎯🗯 Dataset generation for AI chatbots, NLP tasks, named entity recognition or text classification models using a simple DSL!

Overview: Chatito is a tool designed to help developers generate datasets for training chatbot models, natural language processing tasks, and named entity recognition using a simple DSL. It provides an online IDE, DSL specification, AST parser, and a TypeScript npm package for easy integration. Key features: - Online chatito IDE - Chatito DSL specification - DSL AST parser in pegjs format - Generator implemented in TypeScript + npm package Tech stack: TypeScript Why it matters: Chatito simplifies the process of generating datasets for AI chatbots and NLP tasks, making it easier for developers to train and validate their models without needing to manually create complex datasets.
Python 889 stars

topoteretes/awesome-ai-memory

A list of AI memory projects

Overview: This repository is a curated list of AI memory projects, categorized by open-source projects, closed-source projects, and companies, as well as by storage types (graph, vector, both). Key features: - Browse examples in the Cognee QuickStart - Contact us at info@topoteretes.com or join us on Discord - Follow us on Twitter Tech stack: Python Why it matters: This list provides a comprehensive overview of AI memory tools, helping developers and AI practitioners choose the best tools for their projects. Repository: topoteretes/awesome-ai-memory GitHub stars: 889 GitHub description: A list of AI memory projects
Rust 889 stars

vinhnx/VTCode

VT Code is an open-source Rust terminal coding agent.

### Overview VT Code is an open-source Rust terminal coding agent designed for efficient and auditable code management. It allows you to explore, plan, edit, and verify your code with every step auditable. ### Key Features - **Planning**: Plan read-only workflows. - **Safety**: Auditable command policy and sandboxing. - **Self-correct**: HOVER loop on tool mismatch. - **Long runs**: Headless execution with compaction and resumption. - **Integrations**: MCP, Skills, plugins, and editor bridges. - **Models**: Hosted or local providers per task. ### Tech Stack Rust 1.98.1+ (edition 2024) is used for the core functionality. ### Why It Matters VT Code ensures every step of your code changes is auditable, making it a valuable tool for teams looking to improve collaboration and maintainability. ### Repository Repository: [vinhnx/VTCode](https://github.com/vinhnx/VTCode) GitHub Stars: 887 GitHub Description: VT Code is an open-source Rust terminal coding agent.
Python 888 stars

OpenMOSS/AnyGPT

A unified multimodal language model based on discrete sequence modeling

Overview: AnyGPT is a unified multimodal language model that integrates text, speech, images, and music, offering a comprehensive solution for intermodal conversions and multimodal conversations. Key features: - Unified processing of text, speech, images, and music - Base model weights available for inference - Chat model capable of engaging in free multimodal conversations - AnyInstruct dataset for generative models Tech stack: Python Why it matters: AnyGPT leverages discrete sequence modeling to unify and process various modalities, enabling seamless intermodal conversions and enhancing the capabilities of large language models. Repository: OpenMOSS/AnyGPT GitHub stars: 888 GitHub description: A unified multimodal language model based on discrete sequence modeling
Python 887 stars

OpenDCAI/OpenWorldLib

Unified Codebase for Advanced World Models.

### Overview: OpenWorldLib is a unified codebase designed to facilitate advanced world modeling and simulation. It provides a comprehensive framework for understanding and predicting the complex world through multimodal interactions and long-term memory capabilities. ### Key Features: - **Unified Codebase**: A single, centralized platform for developing and managing world models. - **Multimodal Interaction**: Supports various types of input and output, including text, images, and sensor data. - **Long-Term Memory**: Stores and leverages historical data for informed decision-making. - **Simulation Capabilities**: Enables realistic simulations of various environments and scenarios. ### Tech Stack: - Python ### Why It Matters: OpenWorldLib serves as a foundational tool for researchers and developers looking to create sophisticated world models and simulations. Its unified approach simplifies the development process and enhances the accuracy and efficiency of predictive models, making it invaluable for fields such as artificial intelligence, robotics, and game development.
Rust 887 stars

metasequoiaime/chinese-ime-lm

面向中文输入法的语言模型：语料、n-gram 统计、字级神经模型，以及判断它们好坏的评测集

Overview: 这个项目是一个面向中文输入法的语言模型工程，包括语料、统计模型、神经模型以及它们的评测集。仓库中有两条并行的路线，一个是基于 n-gram 的统计模型，另一个是基于字级神经模型。两者都使用了相同的评测集进行评估，评测集本身比任何一个模型都更值得先看，因为它提供了"某个改动好不好"的数字衡量标准。 Key features: - n-gram 路线：语料清洗、KenLM 统计、Viterbi 解码 - 神经模型路线：语料获取、字级 Transformer 训练、评测 Tech stack: Rust Why it matters: 通过使用评测集，可以量化模型的好坏，而不仅仅是依赖于模型的性能排名。评测集本身比任何一个模型都更值得先看，因为它提供了"某个改动好不好"的数字衡量标准。