Trending AI projects

A self-hosted developer control center for managing coding agents and automations, allowing users to run multiple AI agents across local, remote, and cloud backends.

Cua is an infrastructure platform that provides AI agents with controlled desktop environments, automation drivers, and evaluation tools to perform tasks across graphical user interfaces and operating systems.

Hi3D v1.0 is a GPU-only 3D geometry refinement tool: it takes a coarse mesh plus a reference image and produces a high-detail mesh through 1–3 progressive refinement passes (up to 1536³ effective resolution). It ships inference code only, requires Linux/Python 3.10/NVIDIA GPU with CUDA 12.x, and is licensed under a custom agreement permitting commercial use with branding and scale-based license triggers.

Research-Starter-Kit is an open-source Chinese handbook by a Nanjing University professor that teaches research beginners how to do research and write papers, covering everything from finding papers to writing each section and handling rebuttals. It's early-stage, free, and aimed at students entering labs for the first time.

An extension for the pi coding agent that allows users to experiment with and swap system prompts to test instruction-following and prompt robustness.

World Monitor is an open‑source, AI‑enhanced dashboard that aggregates global news, geopolitical, financial, energy and infrastructure data into interactive 3‑D/flat maps. It runs locally with Ollama for LLM‑generated briefs, offers a Tauri desktop app, and provides programmatic access via an MCP streaming server, REST OpenAPI, CLI, and SDKs (Python, Ruby, Go). The project is written in TypeScript, uses Vite, globe.gl, deck.gl, and is licensed under AGPL‑3.0.

Browser Use is an open‑source Python/TypeScript library that gives LLM agents full control over a real or cloud‑hosted web browser. It provides 31 browser actions as tools, supports custom functions, works with OpenAI, Anthropic, Google models or its own BU2 model, and offers three usage modes: a fully hosted cloud API, a CLI for existing agents, and a library you embed in your code. Ideal for automating web tasks, building AI assistants that browse, or researching LLM‑driven web navigation.

nanoMuse is an open-source personal AI agent that runs on your own devices — phone, desktop, and web — and takes real actions like using apps and browsing, while asking before irreversible steps. It offers a free relay allowance or self-hosting, works across many platforms, and keeps your data on hardware you control.

Octop is a self-hosted, multi-user AI assistant platform that provides a private, collaborative environment for managing specialized agents, knowledge bases, and automated workflows across web, desktop, and messaging channels.

OpenSpec is a framework that adds a lightweight specification layer to your codebase, allowing you to align with AI coding assistants on project requirements and plans before code is generated.

A contract-driven regression testing tool for RPC services that automatically composes test chains and verifies behavior against approved baselines to detect regressions.

A collection of tools, plugins, and skills that enable AI coding agents to securely build, deploy, and manage AWS infrastructure using the Model Context Protocol.

awesome-claude-skills is a curated directory of 1,000+ reusable 'Claude Skills' — instruction packages (a SKILL.md file plus optional scripts) that teach AI agents how to do specific jobs, from editing Word docs to running browser tests. Skills load progressively so an agent can carry hundreds without blowing its context window. The list spans 10 categories and links to external repos, mostly from the community. Composio, the maintainer, also ships an MCP Gateway and a 'connect-apps' plugin so Claude can perform real actions across 1,000+ integrations (email, Slack, GitHub, etc.), with a 3-step quickstart in the README.

Awesome Claude Code is a curated list of resources (guides, plugins, skills, tools) for Anthropic's AI coding assistant, Claude Code. It's organized by category, includes official and community items, and is currently in a transitional state where some older entries are temporarily absent but will be restored.

OpenClaw is an open-source, self-hosted AI assistant gateway that connects to over 20 messaging platforms while keeping your data and credentials on your own hardware.

A desktop app that records and transcribes meetings in real-time, providing live AI coaching, conversation metrics, and automated post-meeting summaries.

AI Startup Skills is a bundle of 7 Agent Skills plus a router for solo founders building AI products in China, covering validation → business plan → policy → pitch → growth → pricing → registration. The router picks the right step based on your progress, outputs flow between steps, and it's installable as one package or per-step. It's informational only, not legal/tax advice, and targets mainland China for policy and platform parts.

An open-source, AI-powered video clipping tool that automatically identifies, edits, and formats viral-worthy clips from long-form videos while offering self-hosting capabilities to avoid platform lock-in.

The OpenAI CLI is an official command-line tool that allows users to interact with the OpenAI REST API directly from their terminal, enabling tasks like image generation, file management, and batch processing without writing custom code.

A collection of agent-based engineering skills and playbooks designed to help AI agents write high-quality, verifiable code by automating rigorous development workflows.

A local-first framework for building and deploying personal AI agents on personal devices, reducing cloud dependency by optimizing for on-device hardware constraints.

huashu-motion is an Agent Skill that lets LLM agents generate, edit, and export 30‑second product promo videos entirely via code (GSAP/Canvas/Three.js) and locally‑rendered music, with an interactive browser editor for point‑and‑click revisions.

An open‑source AI Skill that turns a simple topic into a full stick‑figure video plan—script, storyboard, and ready‑to‑use Gemini Omni Flash prompts—across four curated visual styles, enabling fast creation of educational or story shorts for platforms like YouTube, TikTok, and Instagram.

冷咖啡是一款桌面工作台，通过一键修改多款 AI 编码助手（如 Codex、Claude Code、Grok 等）的配置文件，注入预设的提示词和技能库，并支持一键卸载还原，同时内置工具箱和 API 中转功能。

A system that lets an AI agent control a real iPhone using an ESP32-S3 board as a USB HID bridge and AirPlay for screen mirroring, requiring no developer account or on-device software.

An open-source AI video editing tool that lets users edit footage via natural language chat with AI agents, automating cuts, filler word removal, and subtitles.

Official AI skills for GSAP that teach AI coding agents how to correctly use the GreenSock Animation Platform for web animations.

Apple Design Skill is an AI‑agent plug‑in that lets coding assistants review UI code (Flutter, React Native, Tauri, Electron, SwiftUI, etc.) against Apple’s Human Interface Guidelines. It ships 123 Markdown copies of the HIG, a Node script to refresh them, and a skill definition that produces structured reviews (principles, accessibility, platform conventions, visual design, interaction, content) with citations and framework‑specific fixes. Install via the generic `npx skills add` CLI or by cloning into the agent’s skill folder.

VectorCraft is a high-performance, open-source vector illustration application built in Rust that features native support for AI agents via the Model Context Protocol.

Open Generative AI is a free, open-source AI studio for generating images, videos, and lip-synced content with 200+ models. It offers both cloud-based access via MuAPI and local inference options, with no content filters or subscription fees.

A collection of AI agent skills that automate social media research, competitive analysis, and content strategy by turning public social data into structured business reports.

OOMWOO is an open‑source, ROS 2‑based robot vacuum you can build yourself. It uses a Raspberry Pi, 2‑D LiDAR, and 3‑D‑printed chassis, runs locally with Home Assistant integration, and is currently available as a Gazebo simulation while hardware files are being finalized.

A curated list of AI agent skills for short drama and manju production, covering scriptwriting, storyboards, video prompts, and more, with test results and installation instructions.

DSH-IM is a plugin for DeepSeek Harness that connects IM bots from WeChat, Feishu, DingTalk, Slack, Telegram, Discord, WhatsApp, QQ, and more to your local Harness instance. It lets you chat with your AI agents from your favorite messaging apps, with full support for streaming replies, file uploads, group histories, and per-bot configurations.

A complete PyTorch implementation of LLMs from scratch, covering the entire pipeline from raw text tokenization and pretraining to advanced RL alignment (PPO, DPO, GRPO).

An open-source context database for AI agents that organizes knowledge, memory, and skills as a virtual filesystem for transparent retrieval and management.

Coucou is an open‑source desktop and mobile app that lives in the screen notch (or a thin top bar) and serves as a real‑time dashboard for AI‑coding agents (Claude Code, Cursor, Gemini, etc.). It shows live diffs, lets you approve/deny permissions, chat with multiple LLM providers, and offers voice dictation, call transcription, music control, crypto portfolio, and more. All data stays on your device; keys are stored in the OS keychain. Available for macOS 15+, Windows 10/11, Linux, iOS 18+, and Android 11+.

A 17‑lesson tutorial that teaches you how to build the surrounding “harness” code for an LLM‑based coding agent (Claude). It walks through adding tools, permissions, context management, memory, task scheduling, multi‑agent coordination, and goal‑driven termination, all around a simple agent‑loop that lets the model decide when to call tools.

Trader‑skill is a set of Claude‑compatible Python plugins that let you copy public investors, monitor news, manage a budget, check risk, and manually place paper‑trades—all using free public data (SEC filings, congressional disclosures, news feeds).

A collection of agent skills that enable AI agents to Claude Code, Codex, and Open Code to create and edit content within Obsidian vaults using specialized syntax and formats.

JiuwenSwarm is a multi-agent orchestration system that enables teams of AI agents to collaborate on complex tasks, featuring autonomous skill evolution and distributed execution.

New API is an open‑source, self‑hosted gateway that unifies access to many LLM providers behind a single OpenAI‑compatible API. It handles routing, usage/cost tracking, fine‑grained access control, and offers a multilingual web console plus extensible JavaScript task plugins. Deployable via Docker or Docker‑Compose, it’s built with Go (backend) and React/Bun (frontend) and is licensed under AGPL‑v3.

Playwright is a comprehensive framework for browser automation and testing that provides a unified API for Chromium, Firefox, and WebKit, now extended with specialized tools for AI agents to interact with the web.

BaoCut is an AI-powered desktop video editor where an AI agent transcribes, subtitles, dubs, cuts, and animates videos directly into an editable timeline, instead of producing one-shot MP4 renders. It runs local or cloud models, exposes the same runtime via desktop app, CLI, MCP, and web, and lets you bring your own agent (Claude Code, Codex) and models.

Next‑Concept Modeling (NCM) is a two‑stage, parallel text‑generation framework that uses a multiscale VQ tokenizer and a transformer to generate sequences coarse‑to‑fine, achieving 9‑31× speed‑ups while matching diffusion baselines on summarisation, translation, language modelling and Sudoku. The repository includes installation instructions, evaluation scripts, and a collection of pretrained tokenizers and generators for several datasets.

An AI agent skill for generating professional single-file HTML presentations with built-in presenter mode and two distinct high-end visual styles.

Crawl4AI is an open-source web crawler that turns any website into clean, structured Markdown ready for LLMs, RAG pipelines, and AI agents. It supports JavaScript-heavy pages, deep crawling, structured data extraction, and can run locally or via a hosted cloud.

OfficeCLI is a self‑contained CLI tool that lets AI agents and developers create, read, edit, and render Word, Excel, and PowerPoint files without Microsoft Office. It offers JSON‑based document introspection, a built‑in high‑fidelity HTML/PNG renderer for visual feedback, and full‑feature support for editing styles, formulas, charts, shapes, tables, and more.