Trending AI projects

A library of 22 packaged AI-agent workflows ("skills") for academic research — paper reading, Nature-style writing, figure generation, reviewer responses, literature search, and patent drafting — installable into Claude Code, Codex, Chatbox, and other agents.

A plug‑in (Agent Skill) that lets Claude, Gemini, Codex, Cursor, etc., act as a full‑featured logo designer. It follows a brief → research → concept → audit → checkpoint → brand‑kit workflow, using a Python toolset and a curated library of 1 400+ SVG logos. Install via Claude’s marketplace or by copying the skill folder, then ask the agent for a logo and receive production‑ready SVGs, favicons, mock‑up boards and a style guide.

A local tool that provides a unified interface to manage, inject, and hot-swap system prompts (personas) for various coding agents on your machine without modifying their core binaries.

An open-source Postgres development platform that provides a Firebase-like experience with built-in authentication, auto-generated APIs, and real-time database subscriptions.

OpenMausBot is an open‑source desktop app that lets you run a team of AI agents (Claude, Codex, Grok, etc.) in a chat‑style interface. It’s a local‑first alternative to Grok Bot, offering per‑bot models, cloud or local computers, approval cards, app integrations via Composio, voice output, and team sharing, all packaged for macOS, Windows and Ubuntu under Apache 2.0.

An open-source prompt library and platform providing a curated collection of prompts for AI chat models, including educational tools and self-hosting options.

XGO-Duck is an open-source, 3D-printable bipedal robot platform designed to bridge the gap between physical hardware construction and reinforcement learning-based motion control.

MarkItDown is a Python utility that converts PDFs, Office files, images, audio, HTML, and more into Markdown text. It is built explicitly for feeding LLMs and text-analysis pipelines — not for pretty human-facing documents. It runs locally (free), offers optional cloud integrations with Azure for higher-quality extraction, supports third-party plugins, and carries a security caveat: treat inputs as untrusted code.

An interactive terminal UI for DeepSeek Harness and other AI coding agents that provides a rich, visual interface for managing sessions, monitoring performance, and interacting with AI models.

Humanizer‑zh is a Claude skill that rewrites Chinese prose to remove filler and repetitive phrasing while preserving facts, certainty, and the author’s voice. Install via `npx skills add …` or by cloning into `~/.claude/skills/`. Invoke with `/humanizer-zh` on inline text or files. It follows a 31‑point rule set (categories A‑F) and includes a test suite. MIT‑licensed.

Edge0 is an open-source inference framework that allows large sparse-MoE models to run on consumer devices by streaming expert weights from storage and using predictive routing to maintain high performance.

Learn Harness Engineering is an open-source course teaching how to build the environment (harness) that makes AI coding agents reliable. It covers five subsystems (instructions, state, verification, scope, session lifecycle), includes 14 lectures, 8 projects, and a resource library of templates. You can start by adding four files (AGENTS.md, init.sh, feature_list.json, claude-progress.md) to your repo. The course is available in 15 languages and includes recent updates on frontier harness designs, graph engineering, and loop engineering.

The Apify MCP Server is a hosted service that implements the Model Context Protocol, letting LLM agents (Claude, VS Code, etc.) call Apify Actors—ready‑made web‑scraping and automation tools—as native MCP tools. It supports OAuth/standard tokens, a local stdio mode, dynamic tool discovery, and “agentic payments” (AGI token, x402 USDC, or Skyfire). Users connect via `https://mcp.apify.com`, discover actors, and invoke them (e.g., `call‑actor`), enabling AI assistants to fetch live web data or run custom automations without writing HTTP code.

A persistent runtime for coding agents that allows them to run in background terminals across multiple machines with session restoration and status tracking.

Andee is an autonomous Android agent that lives on your device, using accessibility services to interact with any app, manage local memory, and perform tasks without relying on cloud-based vendor assistants.

HumanLayer is an open-source monorepo that lets teams run AI coding agents on their own machines and manage them from a unified desktop/web dashboard, with a daemon on each host, and it sends no data back to the maintainers when built from source.

vLLM is an open‑source library for high‑throughput, low‑latency serving of large language models. It introduces PagedAttention for efficient KV‑cache handling, supports many quantisation formats, integrates fast attention kernels, and offers an OpenAI‑compatible API that runs on GPUs, CPUs and other accelerators.

FilmCraft is a professional, open-source non-linear video editor built in Rust that features native codec support and an automation-first architecture designed to be driven by AI agents.

Scrapling is a comprehensive Python web scraping framework that combines fast HTTP requests, browser automation, and AI-agent integration to simplify data extraction from complex, modern websites.

AI‑Agent‑Book is a multilingual open‑source textbook on building LLM‑based AI agents, paired with 109 runnable Python experiments covering everything from prompt engineering to multi‑agent collaboration.

Codex CLI is a local coding agent from OpenAI that allows developers to to run AI-assisted coding directly from their terminal.

Prism is a dynamic sparse attention framework that enables efficient, high-resolution (2K) training of joint video-audio generation models by focusing computational resources on the most informative visual and audio regions.

An OpenRuntime for AI agents that provides centralized LLM routing, cost attribution, and governance for coding agent harnesses.

A personal AI assistant and Agent OS that provides secure local or cloud deployment, self-evolving memory, and multi-channel connectivity for automation and personal knowledge management.

Earshot is a Python‑only, CPU‑only toolkit that lets AI agents listen to audio (describing, matching, finding events, judging real vs synthetic, generating transcripts) and also create sounds from code via ‘forge’ (search‑guided synthesis) or turn recordings into reusable synthesis recipes via ‘fit’. It provides a CLI, Python API, and MCP server, uses open CLAP and AudioSet models, and includes validation metrics and ready‑made Claude skills.

A collection of official plugins and integrations for the Cursor AI editor that enables agents to interact with external SaaS tools, developer services, and advanced automated workflows.

A system prompt skill for AI coding assistants that forces the model to provide concise, action-oriented, and structured responses by removing conversational filler.

An observability and enforcement framework that hooks into AI agent harnesses to block dangerous tool calls and provide detailed execution traces.

A CLI that discovers, installs, updates, and runs reusable prompt bundles (“skills”) for many AI coding assistants (Claude Code, Cursor, OpenCode, etc.), pulling them from any Git repository or direct URL.

A collection of 49 specialized instructions and an MCP connector that enable Claude to perform automated audits and analysis for Google Ads, Meta Ads, and SEO.

An autonomous, local-first sales agent that automates prospecting and cold email outreach using your existing Claude or ChatGPT subscription to avoid per-token API costs.

A hosted MCP server that exposes Google Ads, Meta Ads, and GA4 APIs as over 250 AI‑callable tools, enabling Claude, ChatGPT, Cursor, n8n, etc., to manage campaigns, creatives, budgets, and analytics via natural‑language prompts. No developer tokens needed; OAuth login and approval‑gated writes are provided through Ryze’s paid service.

Hello-Agents is a free, open-source Chinese tutorial that teaches you how to design and build AI-native agents from scratch, covering theory, classic patterns, framework development, advanced skills like memory and RL, and real-world projects such as a travel assistant and a simulated town.

An MCP server that prevents LLM agents from flooding their context windows with raw data by using sandboxed tools, session indexing, and a code-first analysis paradigm.

A context-compression proxy for AI agents that uses model-driven summaries to reduce token usage and enable month-long coding sessions within small context windows.

Rakazo is an open-source, self-hosted platform for running persistent AI agents that can access local files, browsers, and terminals across desktop and mobile devices.

A suite of Claude Code skills that assists academic research from literature review to publication, with a human-in-the-loop design, citation verification, and style calibration, but without guaranteeing the authenticity of underlying data or experiments.

CLI Proxy API is an open‑source proxy server that presents a unified OpenAI‑compatible API for many LLM providers (OpenAI, Gemini, Claude, Grok, Kimi, Meta, etc.). It handles OAuth logins, multi‑account load‑balancing, streaming, function calls, and multimodal inputs, and offers optional dashboards for usage and cost tracking. Ideal for developers with multiple AI subscriptions who want a single local endpoint for their CLI tools.

Expo Skills is a set of AI‑agent plugins that teach coding assistants (Claude Code, Codex, Cursor, etc.) how to work with the Expo framework. Install via a plugin marketplace or the `skills` CLI, then agents can answer Expo‑specific questions, scaffold code, manage EAS builds, upgrades, OTA updates, and more—leveraging the official Expo docs as the source of truth. The repo is MIT‑licensed and maintained by the Expo team.

huashu‑art‑motion is a Python‑based “skill” that lets AI coding agents generate stylised, code‑driven animation across dozens of art‑style recipes, with built‑in QA, audio synthesis, and explainer‑video grammars.

OpenBot is an open‑source, self‑hosted platform that lets you run AI “coworker” agents inside your own infrastructure. Each Bot gets its own container with a browser and file workspace, all tool calls pass through a policy‑enforced gateway that logs every action. Built on Docker Compose, Bun, PostgreSQL and the AG‑UI protocol, it works with any LLM (OpenAI, Anthropic, Google, or self‑hosted) via CopilotKit Intelligence. Features include per‑Bot computers, CEL‑based access policies, human‑in‑the‑loop hand‑offs, component‑rich responses, audit trails, encrypted credentials, and scheduled routines. The repo is a template (MIT‑licensed, Alpha status) meant to be cloned and customized for enterprise AI assistants, RAG bots, or automated workflows.

Claude Code is a terminal-based coding assistant that understands your codebase and handles tasks via natural language. Install via curl/Homebrew (Mac/Linux) or PowerShell/WinGet (Windows), then run `claude` in your project. Official plugins live in a separate repo; bugs can be reported via `/bug` or GitHub issues; community support is on Discord. Anthropic collects usage data, conversation data, and bug feedback, with stated privacy safeguards.

An LLM-powered trading agent that chats with you over Slack/Discord/Telegram/web, pulls market data, runs backtests, and executes trades across 18 brokers — with a "grounding gate" that verifies every reported number against real tool outputs to prevent AI hallucination.

An AI-powered code review CLI tool that combines deterministic engineering with an AI agent to provide high-precision, token-efficient reviews of Git diffs and full files.

A local-first desktop application for DeepSeek Harness that provides a native host experience, portable agent presets, and editable PPT generation.

A starter kit for animating characters using AI coding agents, combining p5.js and p5.brush to generate stylized 2D videos from text prompts.

A collection of agent skills designed to automate UI engineering tasks like accessibility audits, layout reviews, component stress-testing, and design-to-code implementation.

GEOInflow is a self-hosted analytics platform that tracks AI crawler activity and website traffic, helping teams monitor how AI search engines and training bots interact with their content.