Trending AI projects

LTX‑2 is an open‑source DiT‑based foundation model that generates synchronized audio‑video from text prompts. It offers fast distilled and high‑fidelity DFR pipelines, supports 4K/HDR output, and provides a suite of components (text encoder, video/audio VAEs, up‑scalers, LoRA adapters) with detailed installation and usage instructions.

A Node.js skill that lets AI agents browse, compare, and purchase real products from the Filtmall marketplace, handling discovery, checkout, order tracking, and after‑sales via a JSON‑based CLI.

An open-source AI-powered no-code platform for building business systems that allows AI agents and humans to collaborate on a production-proven infrastructure.

A framework for controllable video generation that uses 3D Three.js code as a structural guide to improve spatial and motion consistency in AI-generated videos.

An open-source SDK for building custom hardware gadgets that connect to the Muse AI ecosystem, supporting both ESP32 microcontrollers and Linux-based devices.

DwarfStar (antirez/ds4) is a native inference engine for running select large language models (DeepSeek V4, GLM 5.x, Qwen 3.8) on consumer hardware (Apple Metal, NVIDIA CUDA, AMD ROCm). It provides CLI, REPL, and HTTP server interfaces, supports SSD streaming, multi‑GPU and cross‑machine tensor parallelism, speculative decoding, and session persistence. Built on llama.cpp/GGML ideas, it’s a fast‑changing beta aimed at local high‑end laptops and workstations.

A versatile, open-source Android image editor that provides batch processing, background removal, and over 500 filter effects without tracking or proprietary dependencies.

Scenario Agent Skills is a collection of 71 modular Agent Skills that enable AI coding assistants (Claude Code, Copilot, etc.) to generate and edit images, video, audio, 3D assets, and brand kits via Scenario’s cloud MCP platform. Install via the `skills` CLI, connect to the Scenario MCP server, and use the skills to automate creative production for games, marketing, and entertainment.

Fish Speech is an open‑source, large‑model‑powered text‑to‑speech system offering expressive, multilingual voice cloning, multi‑speaker scene generation, and on‑the‑fly audio editing. It provides two models – Drama 3 (35 B) for long‑form, high‑fidelity drama production, and S2‑Pro (4 B) for faster, consumer‑GPU deployment – with command‑line, API, and Web UI interfaces, Docker support, and detailed installation guides.

A plugin for Claude Code providing 67 specialized developer skills and structured workflows to automate full-stack engineering tasks.

Droidline is an automation tool that lets you control Android phones from your PC using code or AI agents without needing ADB, USB cables, or root access.

A fast, native, and embeddable coding agent CLI that integrates directly into terminal workflows, CI pipelines, and custom applications.

A packaged desktop distribution of AstrBot that bundles the WebUI and backend runtime for fast local installation and use of ChatUI and plugins.

RoboICL is an in-context learning framework that enables robots to perform tasks by adapting a frozen vision-language model through demonstrations and real-time execution history, eliminating the need for model retraining.

An open-source meta-harness that provides a common orchestration layer for multiple AI agents, enabling cross-device synchronization and cloud sandboxing.

A portable Agent Skill that helps AI agents write less-detectably-AI text by fixing narrative-structure tells (not just word choice), with separate rule sets for fiction and professional documents, voice-skill stacking, Chinese-language calibration, and native plugins for five major AI coding agents.

A data protection SDK that identifies and anonymizes Personally Identifiable Information (PII) in text and images using customizable recognizers.

An AI-powered toolkit for designing, training, and simulating motion for a 22-degree-of-freedom crab robot using reinforcement learning and generative design.

A lightweight, terminal-first AI-native development environment that integrates a GPU-accelerated terminal, code editor, and agentic AI capabilities.

PySR is a high-performance symbolic regression tool that discovers interpretable mathematical equations to fit datasets, providing a transparent alternative to black-box ML models.

BrowserOS is an open‑source Chromium fork offering two products: **BrowserOS neo**, a secondary browser that AI agents control to perform logged‑in web tasks (live dashboard, replayable sessions, one‑click integration with Claude Code, Codex, Cursor, VS Code, etc.), and **BrowserOS**, a regular browser with an AI assistant built into new tabs. Both run locally on macOS, Windows, and Linux, keep all data on‑device, and expose a TypeScript + Rust “agent platform” (MCP server, dashboard UI, Bun backend) for connecting agents. It aims to solve the limitations of headless drivers and cloud browsers by using your real Chrome credentials and supporting any MCP‑compatible AI tool.

A public GitHub board for the Frantic platform, where AI agents can claim paid bounties, deliver work, and receive USDC. The experiment tracks how well autonomous or human‑assisted agents perform real commercial tasks, using the runx runtime for verifiable execution receipts.

higgsfield is an open‑source GPU‑node manager and training framework for huge LLMs. It handles node allocation, ZeRO‑3/FSDP sharding, job queuing, and GitHub‑Actions‑based deployment, letting you train models like LLaMA‑70B across a cluster with a simple Python decorator.

A high-performance vector database and similarity search engine written in Rust, designed for storing and retrieving embeddings with advanced filtering and hybrid search capabilities.

A Model Context Protocol (MCP) server that enables AI assistants to control FreeCAD for 3D modeling, script execution, and FEM analysis.

Sphinx adds an understanding check to agent-assisted development: it pauses a git push and asks you to predict what your change will do, explaining mistakes and re-quizzing until you pass. It works with coding agents like Claude Code and Codex, using a reviewer model to generate open-book questions about behavior and failure modes.

Claude Remotion Skill is an open‑source Anthropic Claude agent skill that lets you describe a video in plain English. Claude then writes Remotion (React + TypeScript) code, renders it, checks the frames, and returns a polished motion‑graphics clip following professional design rules.

A personal AI agent that uses interactive inline controls and live browser handoffs to solve the problem of human-in-the-loop requirements for complex tasks.

A Go-based gateway that unifies OpenCode Zen and Go services into a single API, managing key pools, proxy routing, and protocol translation with a built-in management WebUI.

A verified prompt library and template collection for Seedance AI video generation, providing structured workflows to ensure consistency and quality in AI-generated clips.

A trustworthy evaluation framework for autonomous agents featuring 300 human-verified tasks and a strict Pass^3 metric to ensure consistent model performance.

A mobile-first, durable web app for Pi coding agents that allows for multiplayer collaboration and remote access to local AI agents without a cloud VM.

Redis is a high-performance, in-memory data structure server and vector query engine used for caching, real-time analytics, and GenAI memory.

LoongSage is a production-grade reinforcement learning framework for LLMs that decouples agent logic from training infrastructure, enabling scalable and stable post-training with high-throughput asynchronous execution.

A Python framework that transforms scripts into interactive web apps, enabling data scientists to build and share dashboards and AI chat apps without front-end experience.

A self-improving, self-replicating sovereign AI agent that can independently pay for its own compute and manage its own infrastructure using blockchain wallets.

API Mega List is a GitHub repository containing a large, categorized directory of 11,860 APIs (plus 200 sponsored listings) across 24 categories, updated daily, with links and descriptions for each API.

StudyMate is a Node.js‑based AI tutoring system for math and CS, delivered as a DeepSeek Harness plugin (and also as Antigravity, Codex, or ChatGPT plugins). It builds curriculum roadmaps, renders markdown lessons, saves progress in a local workspace, and can export offline HTML. Install via `npx @yunmiao/studymate@latest install`, then converse with the AI to pick subjects, get lessons, practice, and assess.

garak is an LLM vulnerability scanner that probes generative AI models for weaknesses like prompt injection, toxicity, and data leakage to help developers secure their applications.

Adult Tension is an agent skill for AI hosts that provides a persistent, high-tension interactive narrative experience with complex NPCs and a local database to ensure story consistency and memory.

ComfyUI‑Manager is an extension for the visual AI‑pipeline UI ComfyUI. It adds a “Manager” panel that lets users browse, install, update, enable/disable, and remove custom node packages and model files from a curated registry or local cache. Features include one‑click installs, snapshot‑based environment restore, workflow sharing to external sites, component copy‑&‑paste, a CLI tool for head‑less use, and extensive configuration (security level, network mode, pip/uv handling). It streamlines handling the large ecosystem of community‑made nodes that power Stable‑Diffusion image generation in ComfyUI.

Oxynote is a self-hosted, reviewable workspace that combines live metric charts, API documentation, and technical notes into a single, unified page to prevent documentation drift.

A community‑driven set of AI‑agent “skills” that teach coding assistants how to write better Vue 3 code, covering best practices, router, Pinia, testing, debugging, and composable patterns.

Omnira is an open‑source, self‑hostable platform for running AI agents on any LLM model and any machine. It provides a REST API, TypeScript SDK, and CLI to launch agents, stream events, and handle approvals. Features include model‑agnostic support, built‑in and custom tools, reusable skills, persistent memory, sub‑agents, multi‑tenant RBAC, scheduling, and integrations with Slack, Discord, GitHub, webhooks, and more. You can try it via Omnira Cloud or run it locally with Docker Compose; the Apache 2.0‑licensed code keeps all agent data in your own Postgres.

Pixel Agents is a tool that visualizes AI coding agents (currently Claude Code) as animated pixel-art characters in a tiny office, showing their activity in real time. It runs as a VS Code extension or a standalone CLI, and is designed to be agent-agnostic so new AI tools can be added as single subdirectories of code.

GitNexus is a tool that parses your entire codebase into a knowledge graph and plugs that graph directly into AI coding agents via the Model Context Protocol. It precomputes dependency maps, call chains, and impact scores so tools like Cursor or Claude Code get complete structural context in a single query — reducing broken builds and enabling even small models to work on large repos. It offers a CLI + MCP setup for daily dev work and a Web UI for quick browser-based exploration.

NVIDIA Cosmos is an open platform providing omnimodal world models and tools for developing physical AI applications, including robotics, autonomous driving, and synthetic data generation.

A distributed LLM inference system that pools GPUs and memory across multiple machines to serve large models via an OpenAI-compatible API.