Trending AI projects

A video direction skill for Dreamina/Jimeng Seedance 2.5 that transforms creative ideas and reference images into structured video scripts and optimized prompts.

A user-friendly AI agent desktop application that provides personalized assistants with memory, personality, and the ability to operate a computer and browse the web.

LLM Compressor is a Python library that quantizes and compresses large language models for fast, low‑memory inference with vLLM. It supports many low‑bit formats (FP8, FP4, INT4), a variety of algorithms (GPTQ, AWQ, AutoRound, REAP), and can handle multi‑trillion‑parameter models via disk off‑loading and DDP. The library integrates directly with Hugging Face checkpoints, saves models in the `compressed‑tensors` format, and provides ready‑to‑use examples and pre‑compressed checkpoints.

A scalable platform for running AI agent environments using Firecracker microVMs, featuring rapid snapshot-based booting and high-density resource management.

A comprehensive, open-source library of over 600 AI-ready workflows designed to help Amazon sellers automate tasks like advertising, product selection, and inventory management.

A question-driven knowledge repository offering 100 carefully selected AI Agent interview questions across 11 modules, with clear explanations, code examples, and learning paths to help developers systematically master Agent technology and ace interviews.

An MCP server that enables AI agents to read and manage UniFi network, security, and camera infrastructure through a standardized, secure interface.

A design skill for AI coding agents that prevents generic AI-generated UIs by using computational palette generation, contrast verification, and persistent design memory.

A Google Cloud repository full of notebooks, code samples, and apps for building generative AI workflows — covering Gemini models, search, RAG, vision, audio, and setup guides. It's a community resource (not an official Google product) with many linked companion repositories.

LLM Wiki is an open‑source tool that turns a folder of PDFs, notes, slides, and web clippings into a self‑maintaining personal Wikipedia. Claude (via MCP) reads new material on a nightly routine and writes/updates markdown pages, while a Next.js web UI lets you browse, search, and visualize the growing knowledge base.

An open-source directory and scoring system for AI agent tools, MCP servers, and Claude Skills that helps developers find high-quality, security-graded agent capabilities.

JevHarness is a framework that lets developers offload repetitive decision-making to a fast judgment engine (Jev) while using an LLM to author and optimize the underlying task-specific logic.

An MCP server that enables AI agents to perform professional database administration, including index tuning, health checks, and query optimization for PostgreSQL.

Crit is a local, single‑binary UI that lets you review AI‑generated plans, code diffs, or live web apps. You can comment inline, have LLM agents respond and edit code, generate high‑level “story” overviews, proxy running servers, and sync comments with GitHub or GitLab PRs/MRs. All data stays on your machine unless you explicitly share it.

mistral.rs is a Rust‑based engine for fast, flexible local LLM inference. It supports many model formats (Hugging Face, GGUF), automatic hardware‑aware quantization, multimodal inputs, and built‑in agentic features (tool calling, Python/shell execution, web search). It can serve OpenAI‑compatible and Anthropic‑compatible APIs, offers a web UI, and provides Rust and Python SDKs for embedding.

A library of 76 product management and engineering skills as markdown files that give AI agents specialized knowledge and workflows based on Lenny's Podcast and Newsletter.

x-reader is a Python CLI and Agent Skill that fetches the true source behind URLs (tweets, videos, articles, etc.) and reports whether the full content was read, partially read, or failed, providing machine‑readable evidence for AI agents.

An AI-powered research and writing tool that generates publication-ready scientific papers, grant proposals, and reports backed by real-time literature search and verified citations.

A cross-platform voice dictation and AI assistant that allows users to type with their voice in any app and engage in spoken conversations with an AI.

A toolkit for AI agents to detect and remove bureaucratic language, templates, and robotic phrasing from Russian text while preserving the original meaning.

Rome is an MIT-licensed, self-hostable platform for persistent AI agents. Its core idea: agents don't just get smarter — they build and keep their own tools, workflows, and purpose-built apps, so every interaction makes the next one more capable. You can run it with one Docker script or from source, interact through purpose-built apps (not just chat), and Rome can even build those apps for you from a plain-language description. State is ordinary git-tracked code, so nothing is locked in.

A perception and verification engine that turns video, audio, and screen activity into timestamped evidence for AI agents, featuring a deterministic verification system and a dedicated agent workspace.

A library of 107 Claude Skills covering brand, design, content, SEO, product, growth, dev, and ops. Install via Claude marketplace; skills are stack‑agnostic, composable, and include templates/checklists. Provides end‑to‑end brand‑website workflow with live showcases and an Ahrefs‑powered SEO audit suite.

A high-performance real-time inference engine optimized for small-batch, latency-sensitive AI workloads across NVIDIA, AMD, and Ascend hardware.

A readable, terminal-based coding agent that can edit code and run commands, designed as both a tool and a teaching example for building agentic systems.

A local, reversible PII anonymization tool for Italian legal text that allows users to use cloud LLMs without sending sensitive personal data off-device.

AReaL is a scalable, asynchronous reinforcement learning infrastructure designed to train reasoning and agentic models by decoupling training, inference, and agent execution into microservices.

A real-time monitoring platform for Claude Code, Cursor, and Codex that visualizes agent activity, tool usage, and session costs.

PraisonAI is a Python‑first framework for building autonomous AI agents with a five‑layer stack (Prompt, Context, Harness, Loop, Graph) plus managed remote runtimes. It offers SDKs, CLI, UI/dashboard, JavaScript support, and works with 100+ LLM providers for use‑cases like research, coding, content, data pipelines, support bots, and workflow automation.

HelixDB is an Apache‑2.0 graph database written in Rust that natively supports vector (ANN) search and BM25 full‑text search. It lets you store entities, relationships, embeddings, and text together and query them via a unified JSON API or Cypher. SDKs exist for Rust, TypeScript, Python, and Go, and a CLI provides local dev, embedded mode, and a managed Helix Cloud service.

PenEcho is an interactive, AI-powered canvas that allows users to turn sketches and notes into diagrams and working widgets, integrating with AI agents via the Model Context Protocol.

KohakuTerrarium is a Python framework for building AI agents. It provides a reusable runtime with controllers, tools, sub-agents, session persistence, and composition tools, so you can create custom agents quickly without rebuilding the underlying machinery.

LogoCreator is an open‑source web app that generates brand‑ready logos using the FLUX‑2 image model via Together AI. Built with Next.js, TypeScript, Radix, and Tailwind, it runs without an account—just a Together AI key. Features include instant generation, FLUX‑1 edits, PNG/SVG export, optional rate‑limiting and auth, and a history dashboard. The repo provides clear setup steps and lists future tasks such as size selection, cost preview, reference‑logo uploads, and a showcase of redesigned famous logos.

ComfyUI XB‑BOX (XB_ToolBox) is a ComfyUI extension that bundles UI‑friendly nodes for image/video generation, VRAM optimisation, and YuE2 music generation, aimed at helping beginners run large diffusion models on modest hardware.

Giskard is an open-source Python library for testing and evaluating AI agents, providing automated vulnerability scanning and custom evaluation suites to ensure reliability and safety in LLM-based systems.

OpenJev is an open-source decision server that provides fast, structured, and calibrated answers (yes/no, choice, score) from diffusion-based models by reading their internal probability distributions.

A local-first desktop workspace that centralizes the management, execution, and review of multiple AI coding agents and their associated tasks.

A high-performance, geo-distributed federated metadata lake that provides unified access and governance for data and AI assets across multiple sources and regions.

OpenHUTB / hutb is an open‑source, Unreal‑Engine‑based physics simulator for embodied agents (humans, ground vehicles, UAVs, underwater robots). It offers film‑grade rendering, a rich asset library, VR cockpit, extensive sensor suite, Python API, and integrations with CARLA, AirSim, ROS, Nvidia SimReady, SUMO, and more, enabling AI/robotics research on perception, planning, control, reinforcement learning and large‑model pipelines.

A CLI tool for Power BI report automation that allows humans and AI agents to browse, edit, and publish reports using the PBIR folder format.

A Java-based enterprise backend for ESP32 smart hardware, providing a management platform and AI dialogue services with RAG, voiceprint recognition, and multi-LLM support.

Wardrowbe is a self‑hosted web app that lets you photograph your clothes, automatically tags them with an AI vision model, and provides daily outfit suggestions using a text model. It runs via Docker (or Kubernetes), stores data locally in PostgreSQL, and supports multiple users, notifications, analytics, and eight UI languages.

ComfyUI‑H3‑FaceRefine adds custom nodes that detect, crop, re‑generate, and stitch back tiny faces in MiniMax H3 videos, fixing the model’s poor performance on small heads. It offers automatic and manual face‑selection workflows, robust tracking, optional cut detection, and integrates with existing MiniMax H3 pipelines.

A comprehensive, tutorial-driven project that demonstrates how to build an AI-powered data analysis agent using LangGraph, FastAPI, and hybrid search to translate natural language into accurate SQL queries.

An AI-driven cultivation world simulator where every NPC is an independent LLM agent, creating an emergent narrative based on a rigorous set of cultivation rules.

UnitySkills is an open‑source Unity plug‑in that exposes 839 REST‑based “Skills” for AI assistants to read and modify Unity projects. It offers a three‑tier permission system, dry‑run previews, transactional batch calls, audit logs, and per‑skill rollback snapshots. The plug‑in integrates with major AI coding terminals (Claude Code, Antigravity, Codex, Cursor, Kimi Code, OpenCode) via one‑click installers, enabling safe, automated editor workflows.

GPTaku Plugins is a collection of 19 open‑source extensions for Anthropic’s Claude Code IDE. They let you bypass blocked web pages, extract design tokens, get code reviews from a stronger model, generate full product specs, and more—all without leaving the Claude Code session. Installation is done via Claude Code’s `/plugin marketplace add …` command, and most plugins need no extra API keys; a few (Google Workspace, ChatGPT‑Pro review, Codex CLI) require a one‑time credential setup. The suite is MIT‑licensed and aimed at removing the practical “walls” developers hit when building with AI.

A Markdown‑based “skill” that plugs into AI coding assistants (Claude Code, Copilot, Gemini, etc.) to teach them Home Assistant best‑practice patterns for automations, dashboards, backups, AppDaemon, and more. Install via `npx skills add …` or Claude Code’s plugin marketplace; the assistant then automatically applies the guidance when generating Home Assistant configuration.