Trending AI projects

A tutorial‑style, open‑source Python project that shows how to build a modular, agent‑driven Retrieval‑Augmented Generation system with LangGraph, hierarchical chunking, conversation memory, query clarification, multi‑agent map‑reduce, self‑correction, and evaluation via RAGAS. Works with Ollama, OpenAI, Anthropic, Google Gemini, Qdrant vector store, and optional Langfuse observability.

AIKON is a bridge between 2007-era Nokia phones and modern AI models, allowing users to chat with LLMs like Claude and OpenAI on legacy hardware via a custom Go server.

EngineWire is a native Unreal Engine plugin that lets LLM‑based assistants (Claude, Cursor, VS Code, etc.) control the editor through a single MCP tool called `unreal`. It exposes ~400 actions—spawning actors, editing Blueprints, building materials, running Play‑In‑Editor, taking screenshots, and more—via a simple search/describe/execute flow over either a local HTTP server (port 3000) or a Node.js stdio proxy (port 8090). Security is token‑based and localhost‑only by default. Install the plugin, enable the server, add the token to your AI client, and you can issue natural‑language commands that the plugin translates into Unreal‑Editor operations.

A cloud-native vector database that stores objects and vectors to enable semantic search, hybrid search, and retrieval-augmented generation (RAG) at scale.

AutoHunter is an AI‑driven, Docker‑based platform that continuously hunts for web vulnerabilities. It uses LLM‑guided agents (Collector, Worker, Reviewer) together with real tools like nmap, nuclei, and sqlmap. Humans only perform final review. The system supports multiple search engines (FOFA, Quake, Shodan, etc.), offers a Vue dashboard, and persists data in SQLite. It’s licensed Apache 2.0 and intended for authorized security testing.

Chat SDK is a TypeScript library that lets you write a single bot and run it on many chat platforms (Slack, Teams, Discord, etc.) via plug‑in adapters, with built‑in support for AI streaming, rich UI cards, and state management.

An open-source runtime for AI agents that provides hardware-isolated Firecracker microVMs with instant snapshot-based boot, pause, and resume capabilities.

A collection of 53 plain-text AI agent skills that provide specialized expertise in Amazon selling, from keyword research and listing optimization to FBA fee calculations.

GEM-X is a unified framework for estimating 3D body pose, hand pose, and temporal motion from images and video. It provides a single codebase for inference, training, and evaluation, with pretrained models and support for robotics applications.

A family of Conformer-based acoustic models for state-of-the-art speech recognition and emotion detection, with strong support for Russian and 70+ other languages.

fun-rec 是一本开源书籍，系统讲解推荐系统从传统级联架构到生成式范式的技术演进，包含理论、算法和实战项目，目前仍在开发中。

A comprehensive command center for AI agents that integrates chat, memory, skills, and a terminal into a single workspace for orchestrating multi-agent swarms.

An open-source acoustic drone detection system that identifies multirotor drones by the harmonic sound of their propellers rather than radio signals.

A comprehensive skill set for AI assistants that automates the entire MCU firmware development lifecycle, including building, flashing, and debugging across various toolchains.

An AI-enhanced video player for English learners that features local AI subtitle generation, interactive subtitle navigation, and AI-powered sentence analysis.

Archestra is an enterprise-grade AI platform that provides a unified gateway for managing LLM access, agent runtimes, and tool integrations with built-in security, observability, and governance.

An automated vulnerability research pipeline engine that uses YAML-defined workflows to orchestrate data ingestion, filtering, and LLM-based security assessments.

An open-source framework that uses coding agents to transform game requirements into production-ready assets and engine-specific code for UE5, Unity, Godot, Blender, and three.js.

Goldie is a CLI tool that automates the creation of App Store and Google Play screenshots and preview videos by replaying UI flows on simulators/emulators, framing captures with device bezels, applying themes, and outputting store‑ready assets. It integrates with LLM‑powered coding agents via a ready‑made skill, letting agents generate and tweak the assets through natural‑language commands.

Atomic Agents is a Python library for building modular, type‑safe LLM agents. It uses Pydantic schemas, system‑prompt generators, and plug‑in context providers to let developers compose small, reusable components (agents, tools, context providers) into chatbots, RAG systems, web‑search assistants, etc. A CLI (Atomic Forge) supplies ready‑made tools, and the package works with OpenAI and other providers via the Instructor SDK.

agentOS is a lightweight sandbox library (and hosted service) that runs isolated Linux‑like environments inside a single process using V8 isolates and WebAssembly. It offers ultra‑fast cold starts, tiny memory use, fine‑grained permissions, persistent storage options, and a rich TypeScript/HTTP API for running shell, Python, JavaScript, and more—ideal for AI agents, serverless tasks, and secure compute.

SolarWM is a unified, open-source framework for training and running long-horizon video world models, supporting multiple backbones like Wan2.2 and MiniMax-H3 with a shared data and training pipeline.

A high-performance C++20 inference engine that mirrors vLLM's features and throughput but removes the Python/PyTorch dependency for a 140x smaller deployment footprint.

A collection of samples and tutorials for Amazon Bedrock AgentCore, providing the infrastructure to deploy and operate framework- and model-agnostic AI agents securely at scale on AWS.

A set of plugins that enable AI agents like Claude Code and Cursor to edit ChatCut video projects through natural language prompts.

Copilot API is an open‑source Node/Bun server that provides a single local OpenAI‑compatible endpoint, routing coding‑assistant requests to GitHub Copilot, built‑in Codex/xAI, or many third‑party model providers. It supports Chat Completions, Responses, and Anthropic Messages, offers streaming via SSE/WebSocket, includes a desktop Electron GUI, Docker deployment, and a CLI for quick start.

Biomni is a general-purpose biomedical AI agent that automates research tasks and hypothesis generation by combining LLM reasoning, retrieval-augmented planning, and code execution.

An adapter that enables the use of the Claude Agent SDK within Agent Client Protocol (ACP) compatible clients, supporting advanced agent features like subagent sessions and tool call contracts.

SMG is a production-grade model-routing gateway for LLM deployments: it balances traffic across self-hosted inference engines (vLLM, SGLang) and cloud APIs, performance-optimizes GPU use through cache and load-aware routing, and provides enterprise security, observability, and MCP/tool integration behind one unified API.

VCP is an infrastructure framework that transforms AI agents from stateless tools into persistent, autonomous entities with long-term memory and environmental awareness, enabling them to exist as continuous companions rather than reactive assistants.

An automatic model router for Claude Code and OpenAI Codex that directs simple tasks to fast models and complex tasks to strong models based on per-turn analysis.

VNCCS is a ComfyUI-based pipeline for generating consistent visual novel character sprites. It helps users create a single character that looks identical across all images, manage poses, clothes, emotions, and backgrounds, and work through a full character-production cycle using models like Illustrious, Anima, Qwen Image 2.1, Flux Klein9b, and MiniMax H3.

A custom ComfyUI node set for Minimax H3 that provides neural latent upscaling and tiled re-sampling to accelerate high-resolution video generation while avoiding VAE round-trip overhead.

An industrial AI workbench for creating motion comics and short dramas that solves character consistency through a structured asset-to-keyframe workflow.

A library for building generative agent-based models that simulate social interactions in grounded environments using a Game Master pattern to resolve agent actions.

A Model Context Protocol (MCP) server that connects Garmin Connect to AI assistants, enabling users to query fitness data, manage workouts, and analyze performance metrics via natural language.

Aether is an open‑source, locally‑run AI assistant for Android, iOS and macOS. It provides a ChatGPT‑style UI, uses the Pi framework to talk to many LLM providers, includes an Alpine Linux VM for running shell tools, and supports a rich extension system (web search, model‑context gateway, autonomous sub‑agents, etc.). Installable via App Store, APK, or macOS download, it is GPL‑3.0 licensed and welcomes community contributions.

Agent Standup visualises Claude Code, Codex, Mistral Vibe and Cowork AI‑coding sessions as a live‑call dashboard, showing token usage, tool‑call status, real terminal/file output, sub‑agent hierarchy, replay, video export and optional voice/translation, all from local session files.

A security scanning tool that discovers and audits local AI agent components, MCP servers, and skills for vulnerabilities like prompt injection and malicious code.

GitHub Copilot CLI is a terminal-based AI assistant that lets developers build, debug, and refactor code while interacting directly with their GitHub repositories and workflows.

VibeSkills is a framework that helps AI agents automatically discover, select, and orchestrate the most relevant local “Skills” (reusable tool wrappers) for a given complex task. It clarifies requirements, recommends a workflow level (L or XL), builds a dependency‑aware plan, runs the selected Skills, and performs thorough verification checks. Benchmarked on SkillsBench, it improves task success rates while cutting token usage and tool calls.

A collection of source code and assets for HeyGen's HyperFrames launch videos, providing templates for users to preview, render, and remix programmatic video compositions.

A curated catalog and subscription feed for verified DSH plugins, helping users discover, compare, and install reliable extensions while giving developers a channel to distribute their plugins.

A comprehensive, free Chinese-language tutorial series that teaches users how to effectively use AI coding tools like Codex and Claude Code through practical, step-by-step guides.

A collection of shareable, on-demand agent skills and documentation-driven knowledge bases to help AI agents follow specific technical best practices and frameworks.

Files and templates extracted from a 72-hour live-streamed build where three xAI engineers launched a product using coding agents — including agent rules, 40 documented failures, 69 role definitions, 9 role playbooks, and a PR template that forces proof-before-merge.

A Windows-based desktop AI companion featuring a Live2D character that provides interactive support for chatting, work, coding, and learning through a robust agentic engine.

A set of SKILL-formatted prompts and auditing tools designed to remove robotic patterns from AI-generated Chinese text and mimic human writing styles.