Trending AI projects

A curated registry of design system skill files that provide AI coding agents with specific UI rules, tokens, and accessibility constraints to ensure consistent, high-quality design implementation.

A local safety system for coding agents that prevents the execution of dangerous commands and blocks secret leaks through pre-execution blocking.

A local automation tool that connects AI assistants to Google Flow, allowing you to generate and edit videos by controlling the website's interface directly with your own account credits.

Chandra OCR 2 is an OCR model that converts images and PDFs into structured HTML, Markdown, or JSON while preserving layout, supporting 90+ languages and complex content like handwriting and math. It offers CLI, web app, and a vLLM server for deployment.

Xolotl is a high-performance engine that enables running massive frontier-scale AI models on a single consumer gaming PC by intelligently caching model experts across VRAM, RAM, and NVMe storage.

Kirara AI is a multi-platform chatbot framework that connects mainstream large language models to chat services like QQ, Telegram, Discord, and WeChat. It provides a WebUI for setup, supports plugins, presets, image/voice replies, and an HTTP API, making it easy to deploy a versatile AI assistant.

LoopX is a local-first control plane that provides persistent state management for long-running AI agents, allowing them to maintain goals, evidence, and decision history across multiple sessions and different runtimes.

A long-term memory infrastructure for AI agents that uses deterministic, rule-based cognitive maps stored in Markdown to ensure high performance, zero hallucinations, and full auditability.

Claude Quickstarts is an MIT‑licensed repo offering ready‑to‑run example apps that demonstrate how to build Claude‑powered agents (customer support, finance analysis, browser/computer automation, managed‑agent integrations, autonomous coding, etc.). Each sub‑folder includes its own setup guide, showing best practices for tool use, session handling, and integration with services like Slack, Linear, and Sentry.

DSH 大肥鱼是一个桌面宠物插件，为 DeepSeek Harness (DSH) 提供实时代理状态显示和余额监控。它通过原生置顶窗口展示 AI 代理的思考、执行、完成等状态，并自动查询 API 或账户余额，让开发者无需切换窗口即可掌握工作进度和费用。

A family of multilingual, multimodal translation models that support text, image, and audio inputs, featuring advanced instruction following for terminology, formatting, and syllable-controlled dubbing.

SnapOtter is an open‑source, Docker‑based platform that provides 200+ file‑processing tools (image, video, audio, PDF, generic files) with local AI capabilities such as OCR, background removal, up‑scaling and transcription. It runs entirely on‑premise via a web UI, REST API, or pipelines, offering a privacy‑first replacement for many SaaS converters.

Vibe is a cross‑platform desktop app that transcribes audio/video locally using models like Whisper, supports many languages, offers export formats, summarisation, translation, speaker diarisation, and can be driven via GUI, CLI or an HTTP API.

super-board is an agentic software development framework that automates the lifecycle of coding tasks by routing them through AI-powered build, test, and review lanes.

Microsoft Agent Framework is an open‑source SDK for building production‑grade AI agents and multi‑agent workflows in Python, .NET, and Go. It offers provider‑agnostic LLM access, graph‑based orchestration, middleware, OpenTelemetry observability, declarative YAML agents, skill integration, and easy hosting on Azure or Microsoft Foundry.

FastGen is a PyTorch-based framework that accelerates generative AI models by providing a suite of distillation and training tools to reduce the number of steps required for high-quality image, video, and audio generation.

A Chinese‑market‑focused fork of the open‑source TradingAgents framework. It adds seven analyst agents (including policy, hot‑money and lock‑up specialists), adapts trading rules to A‑share specifics (T+1, price limits, lot size), and uses free data sources (mootdx, Tencent, East‑Money, Sina, Tonghuashun). Two tiers of LLMs (quick and deep) drive the pipeline: analysts → bull/bear debate → research manager → trader → three‑way risk debate → portfolio manager, producing a Chinese‑language investment report. Install via `pip install -e .`, configure an LLM provider in `.env`, and run via a Streamlit web UI or CLI. Includes performance‑tracking commands that compute direction‑accuracy against the CSI‑300 benchmark without extra LLM calls.

AutoRAG Agent is a Node‑based, locally‑run RAG system that turns your computer into a searchable knowledge base. It indexes files (PDF, Office, HWP, images, etc.) with BM25 and a built‑in embedding model, supports many external chat sources, and answers natural‑language questions with numbered citations. Available as a CLI, an interactive terminal UI, a desktop Electron app, and a TypeScript library, it emphasizes privacy (no data leaves the machine) and ease of use (no vector‑DB setup).

An LLM-powered platform that converts unstructured documents like PDFs and images into structured JSON data via natural language prompts.

Repomix is a tool that packs an entire codebase into a single AI-friendly file, preserving structure and counting tokens, so developers can easily feed large codebases to LLMs like ChatGPT or Claude for analysis, review, and refactoring.

CosyVoice is an advanced, LLM-based text-to-speech system designed for high-quality, zero-shot multilingual speech synthesis and voice cloning.

AutoGen is an archived Microsoft library for building multi‑agent LLM applications (Python/.NET). It offers Core, AgentChat, and Extensions APIs plus a no‑code Studio UI. Install via pip, run simple “Hello World”, web‑browsing, or multi‑agent examples, but for any production or long‑term project migrate to the actively maintained Microsoft Agent Framework.

Models.dev is an open-source, community-contributed database of AI model specifications, pricing, and capabilities, with an API to access the data. It fills the gap of a single source of truth for AI model information.

A high-performance, multi-language PDF toolkit for text and image extraction and markdown conversion, optimized for RAG pipelines and AI assistants.

A local-first workbench for managing and distributing AI prompts, skills, and programming rules across multiple AI coding tools like Cursor and Claude Code.

An auto-evolution framework that uses AI agents to iteratively optimize the prompts, configurations, and code of AI agent systems and workflows based on failure analysis.

Open‑source Python CLI/library that scores a website’s AI‑visibility (ChatGPT, Perplexity, Gemini, Claude, Google AI) on a 0‑100 scale, then generates concrete fixes (robots.txt, llms.txt, JSON‑LD, meta tags). Includes 16 commands, CI/GitHub Action support, MCP server tools, Astro integration, and a Python API. MIT‑licensed, install via `pip` or zero‑install `uvx`.

PR Lens is an automated tool that generates interactive, animated architecture and data-flow diagrams for pull requests to help developers quickly understand code changes.

Oh My PPT is a desktop Electron‑based app that uses generative AI (text and image models) to create, edit, and export presentation slides. It works locally, supports PPTX import/export with near‑full fidelity, offers a rich HTML editor with drag‑and‑drop, animations, LaTeX, and AI‑generated graphics, and can connect to local Ollama or any OpenAI‑compatible service.

A tool output proxy that intercepts command-line calls from AI agents to provide fabricated or modified responses for red-team testing and guardrail bypass evaluation.

A design-driven framework for generating professional, editable PowerPoint presentations that uses a visual verification loop (PPTX to PNG) to ensure high-quality delivery.

A self-hostable, open-source workspace that integrates chat, task management, and knowledge base tools with AI agents, featuring governed approval workflows and support for the Model Context Protocol.

OneBox is an all-in-one Android utility app that uses an AI agent to control over 100 local tools, enabling users to perform complex tasks through natural language commands while maintaining privacy.

Desktop Commander is an MCP server that enables AI assistants to perform local file management, execute terminal commands, and handle data analysis directly on your computer.

Fabric is an open‑source Go framework that organizes AI prompts into reusable "patterns" for tasks like summarisation, code explanation, and content generation. It provides a unified CLI, a REST API, and a web UI, supporting dozens of LLM providers (OpenAI, Anthropic, Azure, Bedrock, Ollama, etc.) and offering per‑pattern model selection, internationalisation, and extensibility via custom patterns.

Open‑source tool that turns long YouTube videos into viral‑ready vertical Shorts using Whisper transcription, an LLM‑based highlight scorer, and auto‑cropping (API or local mode).

A professional director's console for ComfyUI that streamlines long-form video generation using the MiniMax H3 model through a visual timeline and multi-segment management.

A curated evaluation and recommendation guide for AI API relay services, helping users find stable and cost-effective providers while avoiding common pitfalls like model dilution and unreliable platforms.

easyeda‑agent is an AI‑driven automation tool for EasyEDA Pro. It lets you describe circuit requirements in natural language and have the AI create libraries, schematics, PCB layouts, routing, silk‑screen adjustments and full DRC/DFM checks, writing the results back to the EasyEDA project via the official `eda.*` API.

Meilisearch is an open‑source, fast search engine offering full‑text, semantic and hybrid search via a simple REST API. It can be used as a managed cloud service or self‑hosted binary, supports typo‑tolerance, filtering, faceting, geosearch, multi‑tenancy, and AI integrations (MCP, LangChain, conversational search).

An open-source desktop workbench based on DeepSeek Harness that integrates AI agent capabilities into a native desktop environment with bundled runtimes and a plugin ecosystem.

uniTerm is a cross‑platform desktop terminal that supports 30+ protocols (SSH, RDP, SFTP, databases, Kubernetes, Docker, etc.) and includes an AI Agent capable of autonomously planning and executing multi‑turn shell commands via Anthropic or OpenAI compatible models. Built with Go (Wails v3) and Vue 3, it runs on Windows, macOS, Linux and Android, offers split‑pane workspaces, cloud sync, theming, and is released under Apache 2.0.

An automated lyric motion tool that generates dynamic text-based music videos from lyrics and audio using a library of 860 modular animation components.

MediaPipe is Google‑Google ​​​​​  …

Nango is an open-source platform that simplifies building and managing product integrations by handling authentication, API proxying, and execution logic for over 1,000 services.

FaceFusion is an open‑source command‑line platform for running and managing face‑manipulation jobs (e.g., swaps, filters) with batch, headless, and benchmarking modes, aimed at technically‑savvy users.

A command-line interface that enables humans and AI agents to interact with DingTalk enterprise services like calendars, documents, and tasks through a secure, structured API.

PaperSpine5 is an AI-powered research assistant that automates the entire academic paper writing process, including literature review, drafting, diagram generation, and final document formatting.