Trending AI projects

A C++ framework for 3D reconstruction from images (photogrammetry) that provides libraries and pipelines for solving the Structure from Motion problem.

An orchestration framework for mathematical reasoning agents that uses a verified fact-graph memory to ensure correctness in complex proofs.

An AI-powered e-ink photo frame that uses vision models to score and caption memories, displaying the best photo from "today in history" daily.

A collection of 14 installable skills for AI coding agents that help developers identify requirements gaps and technical unknowns before implementation.

Pizza Bot is a local-first, asynchronous inbox for AI agents that manages long-running tasks, human-in-the-loop approvals, and multi-step workflows across desktop and web interfaces.

A lightweight 2D digital human engine that enables real-time animation and dialogue on mobile browsers and low-power devices without requiring a GPU.

Jac is a unified programming language and toolchain that allows developers and AI agents to build full-stack applications in a single, compiler-checked codebase, eliminating the need for glue code and external database management.

CLI‑Chat‑Bridge is a Node‑JS tool that bridges local AI‑coding CLIs (Codex, Claude Code, OpenCode, Pi) to WeChat, Enterprise WeChat, or Telegram, preserving the native terminal experience and offering a daemon with rich chat commands.

An intelligent inference router for LLMs that optimizes request placement based on load, priority, and KV-cache locality to improve serving performance.

A visual review tool that lets users edit HTML and Markdown files directly and send batch feedback to AI agents instead of using chat prompts.

一份中文省钱订阅指南，教你通过地区切换、礼品卡和特殊脚本以低成本订阅 GPT、Claude 和 Grok，同时附上避坑与封号应对方法。

Founder Playbook distills 15 classic business books (plus one routing meta-skill) into structured, decision-oriented skill files that AI assistants can load on demand — each with decision trees, frameworks, case studies, templates, and honest notes on where the book's advice is outdated. It auto-triggers based on your question in Claude Code and works as paste-in context or reference docs with any other LLM.

ASu Resume Skills provides two Codex plugins: one turns your real experience into a source-mapped, high-density Chinese resume (HTML/PDF/social copy), and the other audits every claim into atomic statements verified against evidence, flagging overstatements without assuming dishonesty.

A curated collection of 69 machine learning interview questions and answers designed to help candidates prepare for technical roles in AI, data science, and engineering.

A curated list of apps and tools that work with OpenRouter, a unified API for accessing hundreds of AI models from major providers. The directory helps users discover applications that integrate with OpenRouter, from coding agents to chat clients, observability tools, and more.

Awesome Claude Prompts is a community collection of copy-paste prompt examples for Anthropic's Claude AI assistant. It covers coding, writing, marketing, career, learning, and business tasks, plus links to official Anthropic resources and newer Claude Skills/Claude Code documentation. The README includes one full example prompt that generates a visual resume as React code.

A curated, personal collection of AI learning resources, including books, courses, tools, and datasets, designed to help developers navigate the field through a practical, project-based approach.

A curated library of reusable short-video prompt patterns for MiniMax H3 and Seedance 2.0 — with an installable Skills set and an Electron viewer that plays real, credited video examples (not just GIF previews).

Open Agent Email is a self‑hosted service that gives each AI agent a real email address and mailbox. It handles inbound SMTP, stores messages, and lets agents read, reply, forward, schedule, or request human approval before sending. Accessible via REST, Python/TypeScript SDKs, CLI, WebSocket, webhooks, and MCP. Features include per‑agent inboxes, thread view, SPF/DKIM/DMARC validation, prompt‑injection screening, approval gate, any SMTP relay, delivery tracking, signed webhooks, and a Next.js dashboard. Built with Python/FastAPI, SQLite/PostgreSQL, and Docker; Apache‑2.0 licensed.

A Go language implementation of the LangChain framework for building composable applications powered by Large Language Models.

A collection of production-grade Spring Boot skills for AI coding agents, preventing hallucinations and enforcing architectural conventions in Java projects.

imajev is an open‑source family of 2 B‑9 B multimodal models that take a photo (or two) plus structured text data and return typed, probabilistic decisions with an explicit *can’t‑tell* flag. Built on Qwen‑3.5 with a LoRA adapter, the models run locally (GPU or Apple‑silicon), are Apache‑2.0 licensed, and achieve top‑ranked scores on JevBench, Image JevBench, and DecisionBench. Ideal for automating high‑volume, well‑defined business decisions such as product‑photo verification, return validation, ticket routing, and safety checks, while deferring uncertain cases to a human.

A macOS assistant that uses AI to control applications in the background via natural language commands, utilizing Accessibility APIs and ChatGPT for planning.

Pooled is an open‑source web app that lets multiple browsers pool their GPUs via WebGPU and WebRTC to run large LLMs (up to 35 B parameters) entirely in the browser. Devices split the model’s layers, exchange tiny hidden‑state packets per token, and the host assembles the final output. It supports speculative decoding, a built‑in coding‑assistant (“Code mode”), and an OpenAI‑compatible Serve API, all with zero‑install, no‑account usage.

WorldCrafter is a research‑grade video generation model that creates camera‑controllable, consistent videos from a single image or a text prompt. It stores an implicit 3‑D‑aware memory so the scene stays coherent across viewpoints. The repo provides a full‑size Base model and a distilled Fast model, scripts for image‑to‑video and text‑to‑video inference, automatic prompt generation via Qwen‑VL, and an open‑source interactive web demo. Setup uses Python 3.11 with PyTorch 2.10 (CUDA 12.8) via uv or conda, and model weights are hosted on Hugging Face.

Open Managed Agents (OpenMA) is an open‑source, self‑hostable platform for running AI agents with any model, harness, sandbox, and cloud provider. It provides durable sessions, tool integrations, and drop‑in compatible APIs for Claude Managed Agents and OpenAI Agents, plus a web console and CLI for configuration.

A UCI-compliant chess engine that uses neural networks to play and evaluate chess positions, supporting a wide range of hardware backends for optimized performance.

An MCP server and CLI tool that provides AI agents with direct access to search performance data from Google, Bing, and analytics platforms to enable evidence-based SEO analysis.

SimFoundry converts real-world videos into physics-ready simulation scenes for robotics, automating object reconstruction and the generation of diverse object variations.

An open-source, Rust-based container and VM runtime for macOS that provides a high-performance alternative to Docker Desktop and OrbStack, including support for AI agent sandboxes.

An open-source machine learning tool that uses active learning to prioritize records in large textual datasets for efficient systematic reviews.

A knowledge and action layer for Claude Code that enables it to write senior-level .NET 10/C# 14 code using Roslyn-powered analysis and architectural enforcement.

A Claude Code plugin that queries multiple AI coding agents in parallelable, shows their answers side-by-side with a synthesis, and supports roles, debate mode, local councils, caching, and background jobs.

A Navidrome plugin that connects to the AudioMuse-AI engine to provide AI-powered features like automatic song mixes, artist similarity, and discovery tools directly within your music player.

draw‑ui is an AI‑powered skill that helps an LLM agent design a UI from a text brief or existing screenshot and optionally output ready‑to‑use HTML/CSS or WeChat mini‑program code.

Pi Hermes Memory is a Pi extension that adds durable, searchable memory (facts, user profile, failures, and reusable skills) across sessions, with secret scanning, background learning, auto‑consolidation, and a two‑tier global/project architecture.

An OBS plugin that provides real-time, local speech-to-text transcription and translation using OpenAI's Whisper, ensuring privacy and no cloud costs.

Tavily MCP is a Model Context Protocol server that exposes four web‑knowledge tools—search, extract, map, and crawl—so LLM‑based agents (Claude Code, Cursor, etc.) can fetch up‑to‑date web data, pull page content, generate site maps, or bulk‑crawl sites. It can be used via Tavily’s hosted endpoint or run locally with a single `npx` command, authenticating with an API key or optional OAuth. Configuration is done through environment variables (`TAVILY_API_KEY`, `DEFAULT_PARAMETERS`, `TAVILY_HUMAN_ID`). The server enables real‑time web access for AI assistants, RAG pipelines, and productivity tools without custom scraping code.

A lightweight multi-agent TUI that coordinates various AI CLI agents like Claude, Codex, and Gemini into a single, controllable workspace with a stable collaboration layer.

An AI-powered research assistant that automates the process of searching, reading, and analyzing academic papers to generate comprehensive, evidence-based literature reviews.

A curated registry of AI agents that implement the Agent Client Protocol (ACP) and support user authentication, providing a discovery mechanism for client applications.

An all-in-one AI creation platform for novel writing and AI role-playing games, featuring integrated AI agents, world-building databases, and version management.

A learned successor to Perlin noise that uses diffusion models to generate infinite, deterministic, and randomly-accessible real-time terrain and climate data.

Torchvision is a foundational library for PyTorch that provides essential datasets, model architectures, and image transformation tools for computer vision tasks.

Triton Inference Server is an open-source AI serving platform that runs models from multiple frameworks on various hardware, offering high performance, flexible APIs, and easy deployment via Docker. It handles batching, streaming, and model pipelines, and can be extended with custom backends.

SHAP (SHapley Additive exPlanations) is a Python library that computes game‑theoretic feature attributions (Shapley values) for any ML model. It offers fast exact methods for tree ensembles, approximations for deep nets, and a model‑agnostic KernelExplainer, plus rich visualisations (waterfall, force, beeswarm, text, image). Install via pip or conda; GPU support is optional. Ideal for interpreting tabular, text, or image models.

BentoML is a Python library that streamlines turning any AI/ML model into a production‑grade inference API. It provides simple decorators to define services, automatically builds versioned artifacts (Bentos), creates Docker images, and offers performance features like dynamic batching and model parallelism. Ideal for engineers who want reproducible, scalable model serving with minimal DevOps overhead.

Embabel Agent Framework is a Java/Kotlin library for building LLM‑augmented, dynamically planned agents on the JVM. It models goals, actions, and conditions, uses GOAP/Utility AI for planning, integrates tightly with Spring, and supports testable, strongly‑typed flows. Quick‑start templates, Maven Central distribution, and extensive docs make it ready for enterprise AI workflows.