Repository index
6792 AI, ML and Web3 projects · page 22 of 68
- Evolving Skills for AI
Collective skill evolution through an agent evolver — let skills develop alongside your AI.
- KVCache-Factory: Cache Compression for LLM
Unified KV-cache compression methods for autoregressive models. Memory savings and faster inference in one package.
- XTREME1 — PLATFORM FOR MULTIMODAL ANNOTATION
Xtreme1 — platform for annotating 3D‑LiDAR, images and LLM data. Supports collaborative work, AI suggestions and export to COCO, KITTI, Hugging Face.
- WLLAMA: LLM in Browser, No Server
WebAssembly binding for llama.cpp allows running large language models directly in the browser.
- PRUNA - ACCELERATING MODELS
Pruna — a Python framework for model optimization that speeds up inference and reduces resource usage without losing accuracy.
- SUGGESTARR: AUTO PICK
Automatically requests movie and series recommendations for Overseerr/Jellyseerr based on viewing history from Plex, Emby, or Jellyfin.
- OPENSOURCE SLACK WITH AI AGENTS
Self-hosted Slack alternative. Teams and AI agents collaborate in channels and DMs. Agents on Claude, GPT, Gemini, Grok with MCP, browser, memory, approvals. Docker Compose + Postgres, AGPL-3.0.
- OPENGENERATIVE UI - EASY GENERATIVE INTERFACES
OpenGenerativeUI is an open-source TypeScript framework for quickly creating generative interfaces with AI support.
- WORD GPT PLUS:AI IN WORD
Word GPT Plus — an add‑in for Microsoft Word that integrates an AI agent directly into the document. It allows generating text, editing and improving content via natural language commands. It supports working with OpenAI, and local models..
- McDonald's Party Planner MCP
CLI tool using Model Context Protocol to organize McDonald's group meals and party bookings.
- FAST COMPILER FOR EXPRESSIONS
FastExpressionCompiler — fast compiler for C# expression trees and a lightweight LightExpression alternative. Provides diagnostic and code‑generation tools for expressions, speeding up work with delegates and closures.
- AI Agent Learning Hub
The project offers a set of learning materials and pathways for building AI agents: from basic concepts to advanced systems with LLMs and external tools.
- NPCPY: LIBRARY FOR NLP AND LLM
The npcpy library simplifies research in NLP, multimodal LLMs, and agent creation by combining ML, knowledge graphs, and more.
- SOURCEY OPEN SOURCE DOCUMENTATION
Sourcey generates static HTML documentation from OpenAPI, MCP, Doxygen, and Markdown. You fully own the code, with no extra dependencies.
- AWESOME-LLM-RAG: BEST RAG SOLUTIONS
A collection of the best resources and libraries for building Retrieval Augmented Generation in large language models.
- AI CODE SAFETY
A hook for AI agents that intercepts dangerous git and filesystem commands before they are executed. Support for 7 popular CLI tools.
- AI VIDEO TOOLS CLAUDE
A set of tools for creating videos with Claude Code. Automating production via AI agents in Python.
- CHIDORI: Reliable Agent Framework
The Rust framework makes each agent run durable, reproducible, and resumable by default.
- SCENEVIEW: 3D SDK Everywhere
3D and AR SDK in Kotlin for Android, iOS, and Web with AI support: llms.txt, MCP server, and rules for Copilot/Cursor.
- AI Model Price Finder
Grandet lets AI agents discover the cheapest API for any model and connect via its Skill and anonymous MCP server, using price data from over 1,800 providers.
- REVERSE ENGINEERING LAB
Open-source project for reverse engineering: 197 articles, MCP tools, and automation of CTF/APK/PE analysis for AI agents.
- KV-Cache Quantization for LLM
KV-cache quantization using Triton kernels and vLLM integration: 3-bit keys and 2-bit values for efficient large language model operation.
- awesome-deepseek-harness - SPEEDING WORK WITH DEEPSEEK
Ecosystem of plugins and tools for DeepSeek Harness: ready-made solutions from dsh-external/hub and the public topic dsh-plugin.
- meta-ads-mcp - META ADS FOR AI
An MCP server for Meta Ads that lets Claude, ChatGPT, Perplexity, and Cursor manage Facebook and Instagram ads without self‑hosting.
- VISION AND LANGUAGE: TOP MODELS
A collection of architectures of well-known vision‑language models with explanations and links.
- CLAUDE WORKFLOW-V2 — UNIVERSAL PLUGIN
Agents, skills, hooks, and commands for Claude Code — all in one. Accelerate development with maximum AI automation.
- DEEPPAPERNOTE PAPER ANALYSIS
Tool for deep reading of a scientific paper and generating high-quality notes in Obsidian style. Compatible with Claude Code, Codex, Cursor, and others.
- AI Comic Generator
Turns scripts into animated videos featuring unique characters, storyboards, and AI-powered video synthesis.
- OpenMed AI for Healthcare
OpenMed is an open-source Python project using AI for medical data analysis and clinical decision support.
- REPOBRAIN — BRAIN FOR YOUR CODE
RepoBrain turns your repository into a smart assistant: ask questions about the code in plain language. It supports Claude Code, Cursor, Codex, Windsurf, and more.
- LANGCHAIN ON RUST: LLM APPS MADE SIMPLE
LangChain for Rust — official port of the popular framework. Write LLM apps in Rust with the same simplicity as Python.
- Desktop AI Agent App
Open-source desktop application for working with AI agents on Windows and macOS. One-click installation of Claude Code, MCP tools, and Skills with sandbox isolation, multi-model support, and Feishu/Slack integration.
- LLM CODE COMPLETION
VS Code extension for LLM-powered code and text autocompletion. 1359 stars on GitHub.
- ANALYSIS CLAUDE CODE
Systematic analysis of Claude Code for designing AI agents. 1,358 stars — an in-depth breakdown of the architecture and design of modern agentic systems.
- Voice input with AI processing
Hold the key, speak, release — AI-enhanced text will appear next to the cursor in any application.
- AI-AGENT FOR SECURITY
0 — open-source AI security agent that autonomously detects, exploits and fixes vulnerabilities in the stack. Research preview from Swiss Applied AI & Cybersecurity Research Lab.
- XALGORIX AI Pentest Agents
Autonomous AI agents for pentesting: reconnaissance, vulnerability discovery, and exploitation in real time. Implemented in Go and TypeScript.
- Opencove — infinite canvas
A workspace canvas for compute agents, tasks, knowledge, and research. Unites Claude Code, Codex, terminals, tasks, and notes.
- AI FOR SIMULINK
A toolkit that equips an AI agent with expertise for efficient work with Simulink and Model-Based Design.
- SMART SECOND BRAIN IN OBSIDIAN
Plugin for Obsidian that connects your private AI‑assistant to your second brain. Lets you ask questions, get answers and save dialogues directly in notes.
- FAST LAUNCH QWEN3.8 ON AMD
The project provides scripts for instant deployment of Qwen3.8-Flash-Next on AMD Strix Halo (gfx1151) GPU via HIP. All dependencies are built automatically, and inference runs with maximum throughput.
- CLAUDE + PR = MAGIC
Run Claude Code in an infinite loop — it creates pull requests, waits for checks, and merges changes.
- Extensible Native Music Player
A lightweight Go music player with a plugin architecture for custom sources, outputs, and UI extensions.
- LEARNCOPILOT: AI Proof Assistant
The project uses large language models as a copilot to assist in proving theorems in Lean.
- GOWRAP - Automating Go Decorators
A convenient CLI tool for instantly generating decorators for Go interfaces.
- PERSONAL-MODEL – CREATE YOUR HUMAN.MD
Create a personal HUMAN.md — a document with skills, goals, and preferences for working with AI agents. Simple template and flexible configuration.
- REVERIFY: VERIFIED REVERSE ENGINEERING
reverify is an AI reverse engineering tool that uses binary analysis tools and verifies model results against the binary, eliminating hallucinations. Written in Python, it integrates with Codex and Claude.
- SAIKU: ONE CUBE FOR ALL
Saiku is an open semantic model that unites the Excel cube, dashboards, and AI agents via MCP based on Mondrian and Calcite.
- OPEN AI FRAMEWORK MERCATO
Framework with ready-made architectural solutions for AI agents: multi-tenancy, RBAC, event stream, pricing, and CRM/ERP.
- HAFT: SMART SOLUTION ENGINE
haft is an engineering solutions engine that tracks their obsolescence and maintains consistency through evidence decay and parity checking.
- POCO-CLAW: Beautiful UI
poco-claw is a beautiful web interface with built‑in IM support, a sandbox, and a Claude Code‑based agent for teamwork.
- SKYBRIDGE — MCP & CHATGPT FRAMEWORK
A full-featured TypeScript framework for building MCP and ChatGPT apps. Type‑safe, React‑based, runs on any platform.
- MULTI-AGENT LLMS SURVEY 2024
A comprehensive review of LLM-based multi-agent systems — progress, challenges, and prospects. The paper has been accepted at IJCAI 2024.
- NEO4J GRAPHRAG — 1189 STARS
Library for building GraphRAG applications in Python with Neo4j. Graphs + RAG = a powerful combination.
- STASHBASE - TURN FILES INTO AI CONTEXT
Stashbase turns local files into a searchable context for AI agents, speeding up work with data.
- Cairn AI State Search
Cairn is an AI-powered state‑space search engine, initially tested for autonomous pentesting. Written in Python, it has already garnered over 1.3k stars.
- McDonald's Coupon Optimizer
Automatically finds the cheapest McDonald's order combinations by combining menu items with all available coupons, saving up to 53% on meals like the Big Mac combo.
- Unified Agent Control Plane
Manage all agents from a single place.
- MCP-WINDBG: AI DEBUG WINDOWS
Model Context Protocol for WinDBG connects AI agents to the debugger, speeding up bug hunting and crash‑dump analysis.
- CROCODDYL - Optimal control for robots
A C++ library for robot control in contact scenarios using differential dynamic programming (DDP) algorithms.
- RN AI Skill Set
Library of ready‑made React Native components optimized for AI assistants. Simplifies creating code‑generation, debugging, and UI skills.
- CHUBBYSKILLS - CHINESE CONTENT KB
13 AI skills for collecting content from Douyin, Bilibili, Xiaohongshu and other Chinese platforms into your personal knowledge base with transcription and MCP server.
- AntSK Offline Agent .NET 9
AntSK is a knowledge base and agent built on .NET 9 with AntBlazor, Semantic Kernel and Kernel Memory, running fully offline. It supports local Large Language Models and integrates with Aspire for monitoring.
- KV Memory Virtualizer
KVMem virtualizes KV memory across backends, enabling ultra‑long‑context LLMs on low‑RAM hardware by transparently paging the cache.
- Meridian Proxy for Claude Max
Meridian is a custom proxy that lets you use your Claude Max subscription with OpenCode, Pi, Droid, Aider, Crush, Cline, and other tools. 1328 stars on GitHub.
- HACKINGBUDDY GPT HELPER
The project helps ethical hackers use LLMs in 50 lines of code or less. Rapid AI integration into pentest workflows.
- AI Chat for Apple Silicon
A local AI chat and agent for Mac with Apple Silicon, running via MLX with Ollama support.
- JUPYTER CONNECTS
MCP server for Jupyter — now AI agents can launch notebooks, run code, and read results directly.
- BEDROCK-CHAT: AWS CHATBOT IN TYPESCRIPT
AWS-native chatbot on Bedrock, written in TypeScript, for quickly creating AI assistants.
- OPENCODE CLOUD AUTHORIZATION
The OpenCode plugin uses your existing Claude Code credentials to start working immediately without a separate login.
- ALPHACLAW - DEPLOY IN MINUTES
Perfect harness for OpenClaw. Launch in minutes, stable for months — no CLI needed.
- EMBEDANYTHING - POWERFUL INFERENCE ON RUST
EmbedAnything is a high‑performance, modular, and memory‑safe framework for inference, ingestion, and indexing on Rust.
- SEAGOAT: LOCAL SEMANTIC SEARCH
SeaGOAT — a local semantic code search engine that works without cloud dependency. It uses embeddings and tree‑sitter to index the repository and search by meaning, not exact match.
- TAU2-Bench: Benchmark for AI Agents
τ-Bench tests AI agents in real‑world tasks. The Python project already has 1.3k stars.
- NEUROAPI: RUSSIAN AI API
NeuroAPI provides secure one‑click installers for Windows/macOS, connecting Codex CLI and Claude Code to the Russian AI API. Works with Anthropic, OpenAI and local LLM models via a unified interface.
- AIRUNNER OFFLINE AI FOR ALL
Offline inference engine for arts, real-time voice dialogues, LLM chatbots, and automated workflows.
- CPA-USAGE-KEEPER: Usage Tracker
Standalone CLI‑proxy usage tracker with SQLite storage and built‑in dashboard.
- AI-Setup: One-Key Sync
Sync AI-agent settings for Claude Code, Cursor, and Codex with a single command. Supports skills, MCP files, and configs.
- ONE LINE — Awwwards‑LEVEL SITE
skills — agent skill that turns a single prompt into a ready‑to‑deploy Astro site of Awwwards quality. It automatically picks design, animations and SEO, then deploys to Vercel or Netlify. Extensible with custom agent‑skills and Codex suggestions.
- MONITORING REDIS AND VALKEY
monitor — TypeScript utility for real‑time monitoring, slowlog analysis and traffic audit in Valkey and Redis. Integrates with Dragonfly, MCP‑server and metrics export to observability systems.
- SKILL BOOST WITH PYTHON
Harness codebase and loop engineer for automating skills with Python.
- OVERSTORY - AI Coder Orchestration
Flexible adapters for Claude Code and Pi. 1300+ stars, TypeScript.
- AUTONOMOUS LLM AGENTS
A collection of research papers on LLM-based autonomous agents. Updated daily — 1311 stars on GitHub.
- mini-AGI continual learning
mini-AGI is a continual learning model, trained from scratch on a laptop with 8 GB VRAM, processing a batch‑1 data stream and capable of incremental knowledge updates.
- Mini‑pi on TypeScript
An ultra‑compact implementation of π in TypeScript (~600 lines) that lets you build your own pi‑agent from scratch for any task.
- AI Dictation Without Keyboard
Open AI assistant for dictation, working locally and offline. Type three times faster without a keyboard.
- AI AGENT FOR EXPLAINER VIDEOS
Explainer videos and product demos are created locally by your AI agent. It uses the Kokoro voice, Whisper word‑timing and a Canvas renderer to turn a script into a narrated MP4. The project is free and open source.
- REACT NATIVE AI - MOBILE APPS
A full-featured TypeScript framework for building cross-platform mobile apps with AI.
- CLIENT CLAUDE CODE IN WEB
Full-featured web client for Claude Code with interactive project management. Built on TypeScript, already 1.2k stars.
- Claude to OpenAI Converter
The project provides an OpenAI Chat API‑compatible layer for calling Anthropic Claude models, allowing existing code to be used without changes.
- PARALLAX: BUILD AI CLUSTER
Parallax lets you quickly deploy a distributed inference service on any infrastructure.
- AI AGENT WEB SCRAPING DATA
The Apify MCP server enables AI agents to retrieve data from social networks, search engines, maps, and e‑commerce stores via ready-made scrapers from Apify Store.
- ClawBench - Benchmark for AI Agents
Open-source benchmark for evaluating browser AI agents on everyday tasks. Already earned ⭐ 385!
- MATLAB for AI Agents
MATLAB Agentic Toolkit unlocks MATLAB capabilities for AI agents, making engineering and scientific processes ready for automation.
- TRACELY-AI - AUTOMATED AGENT TESTING
Tracely-ai automatically turns AI agent failures into regression tests that block PRs and reproduces them in CI for free.
- CONTEXT MEMORY FOR AI AGENTS
Puppyone-cloud is a context engine for AI agents. It provides persistent storage, versioning, and instant retrieval of relevant context for LLMs. Simplifies building agents with long-term memory.
- AGENTDOCK: Managing AI Agents
Secure MCP runtime for running AI agents on local machines, servers, and containers with multi-device orchestration.
- LLM AUTOMATICALLY COLLECTS WIKI
Knowledge compiler: raw sources → interlinked wiki. Inspired by Karpathy's LLM Wiki pattern; TypeScript, ⭐1292.
- PHONECLAW: On-device AI Agent
On-device AI agent for smartphones, powered by Gemma 4, provides fast access to intelligent features without the cloud.
- KUBEAI – Integral Inference Operator
AI Inference Operator for Kubernetes: the simplest way to deploy ML models in production. Supports VLM, LLM, embeddings, and STT.