Trending AI projects

An end-to-end AI agent engineering platform combining a TypeScript framework for building agents and workflows with a production observability and operations console.

This project is an AI-powered e-commerce visual planning tool that converts raw product data into structured, compliant, and high-converting visual assets for designers and AI image generators.

A CLI and MCP server that provides programmatic access to Gemini Notebook, enabling users to automate notebook management and integrate Gemini's RAG capabilities into AI agents and local workflows.

Paper2Agent is a multi‑agent Python system that automatically converts a research paper (and its code) into a runnable MCP server, exposing the paper’s methods as AI‑callable tools. Users install the “paper2agent” skill into a coding assistant (Claude Code, Codex, Gemini CLI), then prompt the assistant to *agentify* a paper URL. The system spawns specialist sub‑agents to parse the paper, set up an isolated environment, wrap the original code, run verification tests, and bundle everything into a ZIP with usage instructions. The resulting server can be run locally or hosted on Hugging Face, and then connected back to the coding assistant for interactive scientific queries.

ChatTTS is a generative text-to-speech model optimized for conversational dialogue, providing natural, expressive speech with control over laughter and pauses.

Applio is a cross‑platform desktop app for high‑quality voice conversion. It provides a Gradio‑based web UI, one‑click installers, optional TensorBoard monitoring, and a plug‑in system for adding new voice models. The tool wraps pretrained speech‑conversion models (e.g., RVC) and is aimed at creators, developers, and researchers who need simple voice‑swap functionality without coding.

A curated directory of open-source AI agent skills designed to help coding assistants write better Swift and Apple-platform code. It provides developers with a central hub to find and integrate expert-level knowledge into their AI coding tools.

A zero-dependency SVG and JavaScript expression engine that provides 32 emotional states for AI assistants, chatbots, and desktop pets.

SRS is a high-performance, AI-driven real-time media server that supports multiple streaming protocols like RTMP and WebRTC across various hardware architectures.

A Nix flake that packages a wide range of AI coding agents and development tools (like Claude Code, Codex, Gemini CLI, etc.) and keeps them updated daily, so you can install and run them with a single nix run command.

A unified framework to evaluate generative language models on a large number of standard academic benchmarks, serving as the backend for the Open LLM Leaderboard.

ContextGem is an open‑source Python library for extracting structured, typed facts from documents using LLMs and low‑cost decision models. You declare aspects (text sections) and concepts (typed values) in plain language; the library auto‑generates prompts, routes each extraction to the appropriate model, validates output, and returns each fact with its source paragraph and optional justification. It supports probability‑based answers from decision models, batch processing, and full documentation.

Rowboat is an open-source, local-first personal AI assistant that continuously indexes your work (email, meetings, chats, notes) into an editable knowledge graph, then uses that context to draft emails, take meeting notes, browse the web, and drive coding agents (Claude Code/Codex). It's fully local and private by default, works with any model, and optionally connects teammates via "Spaces" — shared rooms on a separate open-source server called Harbor, where each person's Rowboat acts only with their own context and posts only what the user chooses to share.

Terminal-Bench-Science is a benchmark for evaluating AI agents on challenging, expert-curated research workflows across the life, physical, earth, mathematical, and engineering sciences. It offers a leaderboard, task dashboard, and a contribution pipeline to measure and advance agent capabilities in scientific research.

不露脸做视频账号的 AI skill 合集：调查长片 + 手绘白板讲解视频，一句选题出片、配音字幕、封面与各平台文案

A webcam-based AI motion capture application for animating 3D avatars (VRM/MMD) in real-time for VTubing and XR purposes.

An end-to-end open source platform for machine learning that provides tools and libraries for building and deploying ML-powered applications.

A curated, language-organized list of machine learning frameworks, libraries, and tools, with extra resources and clear maintenance rules.

Knowledge Catalog is a metadata management platform that creates a dynamic knowledge graph to provide AI agents with the business context and semantics needed to understand structured and unstructured data.

A private, automatic work journal for Mac that uses AI to turn screen activity into a chronological timeline and searchable work history.

An open-source collection of 28 production-grade tutorials for building deployable AI agents, covering everything from orchestration and memory to deployment and security, plus a companion book, YouTube channel, and community resources.

A high-performance CUDA kernel library for LLMs that optimizes GEMM and MoE operations, featuring runtime JIT compilation and extreme throughput on NVIDIA SM90/SM100 GPUs.

A model that predicts spatially consistent 3D geometry, depth maps, and camera poses from single or multiple images using a unified depth-ray representation.

Pi-Harness is a desktop workspace for the Pi Coding Agent that adds visual management, run control, recovery tools, and multi-agent orchestration to the native agent experience.

Llama is a lightweight macOS menu bar application that simplifies running local AI models by automatically configuring them for your hardware and providing a local API for other apps to use.

A multilingual, tiny speech generation model with 0.1B parameters designed for real-time, CPU-based inference and high-fidelity voice cloning.

ZCode Proxy is a local tool that converts Zhipu AI coding subscriptions into standard OpenAI/Anthropic API endpoints, enabling you to use GLM models in any AI-powered coding tool.

HelloAgents is a Python library for building modular multi‑agent applications that use LLM function‑calling. It provides ready‑made agents, a tool registry, context handling, RAG/memory stores, sub‑agents, streaming, and observability, with adapters for OpenAI‑compatible, Anthropic and Gemini services.

Chat On Steroids is a cross‑platform desktop app that lets your own ChatGPT conversation read/write files, run commands, and automate the browser. It uses OpenAI’s MCP plugin API, splits work into persistent “workers”, and provides a loop that sends goals to the model, executes the suggested actions, and feeds back results. Install via the provided binaries, pair a custom MCP app in ChatGPT, load the companion browser extension, then type a task and watch the model code, test, and iterate on your local machine. The project stresses responsible use, does not circumvent OpenAI limits, and is MIT‑licensed.

Nanobrowser is an open‑source Chrome/Edge extension that lets a large‑language‑model act as a user inside your browser. You type tasks in a side‑panel; the agent reads pages, clicks, types, navigates tabs, and can upload/download files. It supports many cloud LLM providers and local models (via Ollama, LM Studio, etc.), offers Ask/Auto modes, per‑site permissions, voice input, memory, and multi‑tasking, all while keeping API keys and page data private. Install from the Chrome Web Store or build from source with Node 24 and pnpm 10.

Hojo-TTS-Light is a collection of ultra-lightweight, bilingual text-to-speech models designed to run efficiently on CPUs and edge devices while maintaining high audio quality.

A tool that converts Chinese stories or images into vertical hand-drawn animations using AI Agents and Remotion, featuring a massive library of 327 artistic styles.

A Claude skill that enables the AI to watch videos from URLs or local files by extracting key frames and transcripts for multimodal analysis.

An AI-powered multimedia tool that integrates speech recognition, translation, and voice cloning for automated dubbing and multilingual content creation.

A starter kit for Claude Code and Codex that turns AI assistants into a personalized AI Operating System through structured context gathering, reliability auditing, and 3D knowledge visualization.

Agent Router is an open‑source, Envoy‑powered control plane that provides a single OpenAI‑compatible API for routing requests to many LLM providers or self‑hosted inference clusters, handling auth, quotas, fail‑over and usage tracking via Kubernetes CRDs and a CLI.

A Model Context Protocol server that gives AI assistants natural language control over Google Workspace services—Gmail, Calendar, Drive, Docs, Sheets, and more—with 120+ tools, multi-user OAuth 2.1, and secure deployment options.

A meta-layer for turning text, articles, or GitHub repos into animated MP4 videos using local coding agents and pluggable HTML-based rendering engines.

A local-first AI e-book reader that uses RAG and semantic search to help users chat with their library and build a private knowledge base.

zeroshot is an open‑source CLI that lets you build and run AI‑agent workflows (coding, review, repair, delivery) expressed as a directed graph. Each step can use a different LLM, and the engine enforces review gates, retry limits, and optional Git/GitHub delivery. Runs are recorded in a SQLite ledger and can be executed locally, in Docker, or via zeroshot Cloud.

Kandev is an open-source control plane for orchestrating multiple AI coding agents in parallel, with kanban task management, multi-agent workflows, and an integrated review workspace. It solves the problem of scaling agent-based development while keeping humans in control of what gets shipped.

A collection of 14 expert skills and an enforcement layer for AI assistants to build, validate, and deploy production-ready n8n workflows using the n8n-mcp MCP server.

GVHMR is a SIGGRAPH‑Asia 2024 research codebase that recovers world‑grounded 3‑D human motion from monocular video. It combines visual odometry (DPVO or SimpleVO) with a gravity‑aligned coordinate system, provides ready‑to‑run demo scripts, Colab/HuggingFace demos, and scripts to reproduce benchmark results on 3DPW, RICH, and EMDB.

A toolkit for running open LLMs on Kaggle's free TPU v5e-8 hardware, providing public OpenAI/Anthropic-compatible API endpoints for coding agents.

Facet is a private, self-hosted photo manager that scores each image across 9 quality dimensions, then offers a web gallery with culling, face recognition, semantic search, sharing links, and optional GPU-accelerated AI — all offline with no cloud accounts.

Mike is an open‑source legal‑AI platform that combines a Next.js web UI, an Express API, Supabase/Postgres, and Cloudflare‑compatible storage to let users chat with, edit, and review legal documents using LLMs (including local Ollama). It includes a Microsoft Word add‑in, connectors for Gmail/Calendar, and reusable workflows, all runnable locally via Docker Compose.

A serving stack that runs GLM-5.3-Flash (125 GB) on a single RTX 3090 24 GB by splitting its 12,096 routed experts across VRAM, RAM, and NVMe, delivering up to ~28 tok/s decode with near-exact output.

ACRYL is an early-stage (v0.2.5, MIT-licensed) "Agentic Development Environment" that keeps your project, context, and work persistent while letting any coding agent (Claude Code, Codex, etc.) plug in as a replaceable worker. It features a canvas-based UI, a catalog of 33 agents, agent-driven app control with per-action approval, a plugin system built on Cordis, shareable "Blend" environments, and three interchangeable surfaces (desktop, local web, terminal). It builds on a pinned, unmodified DeepSeek Harness checkout as its capability substrate.