Repository index
6791 AI, ML and Web3 projects · page 7 of 68
- PROJECT MANAGEMENT SYSTEM FOR AGENT WORK
ccpm — project management system for AI agents based on GitHub Issues and Git worktrees. It launches agents in parallel using isolated work trees. Supports shell scripts and Claude Code.!!!
- DEEP SEARCHER AI AGENT
Deep‑searcher is an open‑source Python framework for deep research of private data. It offers an alternative to Reason and Search with agentic RAG, supporting Claude and DeepSeek models. It has already garnered 8 275 stars on GitHub.
- API-MEGA-LIST - COMPLETE API LIST
A collection of the most useful APIs for quick project start: from automation to full-stack applications.
- Mobile Chat Copilot
Reads QQ, X, and Feishu chats on your phone, suggests reply options, and lets you insert them with one tap.
- Kreuzberg: Documents and AI
Multilingual framework for extracting text, metadata, and structured data from 97+ document formats.
- YAO - All Agents in One
An open service for managing agents and workspaces on all devices: task boards, API, mobile and web interface.
- Prompt engineering framework
The AI framework that adds engineering to prompt engineering (Python/TS/Ruby/Java/C#/Rust/Go compatible).
- LMCache: LLM Acceleration
The fastest KV cache layer for large language models, boosting performance and reducing latency.
- MLX-AUDIO: SPEECH ON APPLE SILICON
mlx-audio — library for TTS, STT and STS on Apple's MLX framework. Runs efficiently on Apple Silicon. 8192 stars on GitHub.
- VIBE CODE WITH PI IN IDE
Vibe‑code with Pi in a lightweight real IDE — AI agent writes, tests and debugs code. Get instant feedback and focus on creativity.
- MobileGestalt Editor for iOS
Edit MobileGestalt keys on iOS 27.x with region-code tweaks, device presets, and backup management.
- BITSANDBYTES: LLM COMPRESSION
Library for k-bit quantization of large language models, enabling LLM execution on regular hardware via PyTorch.
- GPTCache: SEMANTIC CACHE FOR LLM
GPTCache provides semantic caching of requests to large language models, reducing token consumption and speeding up responses. Fully integrated with LangChain and llama_index, it supports Python API.
- GO BACKEND FOR VOICE AND VIDEO AGENTS
Go backend for building text, voice, and video AI agents with SDKs in 7 languages. Agents are defined via YAML configs and folders, with integrated tools: memory, search, telephony, sandboxes.
- Claude Code Tips
45 tips for using Claude Code effectively, from basic techniques to advanced — custom status line, shortened prompt, Gemini CLI as an assistant, and running in a container.
- GITHUB COPILOT IN XCODE NATIVELY
First native Xcode extension integrating GitHub Copilot, Codeium, and ChatGPT directly into the code editor. Works via Source Editor Extension API without wrappers. 8200+ stars on GitHub.
- iOS MobileGestalt Configuration Tool
Edit MobileGestalt keys on iOS 27.x with region-code tweaks, device presets, and backup support.
- SenseVoice: Voice Understanding Model
Multilingual model for analyzing and understanding speech in different languages with high accuracy.
- VIBE: Transcribe Yourself
Vibe — a TypeScript tool for self‑transcribing audio. Upload files, get accurate text without servers.
- NEXA-SDK AI Everywhere Always
Run cutting‑edge LLMs and VLMs on GPU, NPU, and CPU — from PC to mobile and IoT devices. Support for models like GPT‑OSS, Granite‑4, Qwen‑3‑VL, and more.
- TREG: Tool for Agents
treg is an OpenRouter for agent tools, allowing easy integration of LLM into any workflow.
- Video editor with AI
Edit videos using AI agents in Python.
- SOURCERY: SWIFT METAPROGRAMMING
Automate Swift boilerplate code generation, saving hours of development.
- WEBINY-JS: Self‑hosted CMS on AWS
Open‑source TypeScript CMS running serverless on AWS Lambda, DynamoDB, and S3. Supports multi‑tenancy, GraphQL, and an AI‑assistant via MCP.
- PI-DESKTOP: LOCAL AI CODER
Local AI coding agent: Electron + Rust core + OpenAI, Claude and local model support. Fast repository indexing via tree-sitter, plugins and i18n.
- Claude Code Plugin
Replaces the compaction summary with Jev solutions, evaluating each tool call in a single fast request and discarding or truncating stale data while leaving the rest unchanged.
- ADK-GO: AI Agent Framework
An open Go framework for building, evaluating, and deploying complex AI agents with flexibility and control.
- AI engineering from scratch
Learn, build it, and deliver it to others.
- CHIC MCP FOR HOME ASSISTANT
Unofficial but cool MCP server for Home Assistant — 3500+ stars and pure Python.
- LMDeploy: Easy LLM Deployment
A toolkit for compressing, deploying, and serving LLMs. ⭐ 7873 | Python
- Autonomous AI Agents
A framework for building autonomous AI agents in Python with support for multiple LLMs, sandboxes, and advanced architectures.
- K8SGPT: KUBERNETES ASSISTANT
k8sgpt uses AI to analyze Kubernetes clusters, identifying problems and offering instant fixes.
- ODS - YOUR AI SERVER
ODS turns any computer into a powerful AI server: LLM inference, chat, voice, agents, RAG, and image generation.
- LOCAL-FIRST DESKTOP WORKSPACE FOR AI AGENTS
Cross-platform desktop app for managing persistent AI agents (Codex, Claude, Grok, etc.) with isolated workspaces, task queues, and an embedded browser.
- WHATSAPP HTTP API IN MINUTES
WAHA — REST API for WhatsApp with four engines: WEBJS, NOWEB, GOWS, and WPP. Configures in one click, supports MCP server for AI agents. 7.4k stars on GitHub.
- Reef: continuous learning
Infrastructure for continuous learning of self-improving agents based on Python.
- Native Zelda on Apple Silicon
Runs Breath of the Wild and Tears of the Kingdom natively on Apple Silicon by translating NVN API calls to Metal. Requires own game files.
- Auto-Tuning ML Inference Engine
Open-source inference engine that auto-tunes kernels per hardware for 2x faster model execution than Ollama across Apple Silicon, NVIDIA, AMD, Intel, and CPU.
- Claude Day Trading Skills
11 Claude skills for day trading: scan charts, track news, time trades, and replay outcomes before risking capital.
- Self-Hosted A-Share Quant Workbench
A zero-maintenance, self-hosted platform for building, monitoring, and backtesting custom A-share trading strategies with LLM-assisted analysis.
- CODE TO PROMPT INSTANTLY
A Rust CLI tool that converts an entire codebase into a single LLM prompt, including a file tree, templates, and token count.
- VISUAL GUIDE TO TRANSFORMERS
An interactive visualization that reveals how Transformer models work. Easy to understand even for beginners.
- MCP FOR GOOGLE ADS AND META ADS
Single MCP server unifies Google Ads, Meta Ads, GA4, and Search Console. 250+ tools, OAuth no keys, approval-gated writes, free for Claude, ChatGPT, Cursor, and n8n.
- AI Infrastructure Book
Open‑source Chinese book on AI infrastructure, covering LLM inference and training system design with quantitative analysis, plus PDF, code tools, and experiments.
- TUI Plugin for DSH Claude Code
A TUI plugin officially presented in the DSH public account, adding a Claude Code–style interface with a whale top bar, live status, streaming thoughts, double-Esc rollback, and a context + TPS indicator. Installs with a single npm command.
- 48 FREE CLAUDE SKILLS
48 free Claude skills for marketing: Google Ads, Meta Ads, SEO and AI visibility. Simple way to add Ryze MCP via claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp. 1.2k stars on GitHub.
- DRAMACLAW — SCRIPT TO FILM
Universal AIGC video engine: from script to finished clip in one pipeline.
- CHATGPT PLANS, CODEX EXECUTES
ChatGPT acts as the project's brain: it analyzes requirements and forms a plan. Codex turns the plan into working code, preserving context via Model Context Protocol. The result is seamless work of the AI agent without manual switching.
- FRAMEWORK FOR AI AGENTS
Agent-native — a TypeScript framework for creating AI agents. Supports OpenAI, Claude and RAG.
- Hand-Drawn Prompt Gallery
An HTML gallery of numbered hand-drawn style prompts with bilingual text and a reusable style skill.
- RAG Chatbot on Azure
Application template with Retrieval-Augmented Generation on Azure: Azure AI Search + Azure OpenAI for ChatGPT-like interfaces and Q&A.
- CONNECT ANY AGENTS TO LLM
Connect any AI agents and tools to any LLM provider via a unified gateway. Supports Anthropic, OpenAI and other APIs, tracking cost and usage.
- Flexible AI-agent framework
Framework for managing multiple AI agents and organizing complex dialogues.
- refly - Build Agent Skills
The first open-source framework for creating agent skills via visual workflows. Works with Claude Code, Cursor, Codex, and other IDEs.
- iOS MobileGestalt Editor
Edit MobileGestalt keys on iOS 27.x with region-code tweaks, device presets, and backup support in 7 languages.
- Prompt Engineering in Practice
22 prompt engineering techniques using Jupyter Notebook, covering basics to advanced strategies for working with LLM.
- Minimalist Poster Generator
The project provides a Codex skill for creating quiet minimalist editorial posters in a zine style, generating ready-made prompts and images.
- Local Deep Search
Achieves ~95% accuracy on SimpleQA, works with any local and cloud LLMs and more than 10 search engines, all encrypted and local.
- Terminal AI Agent Management
Manage multiple terminal AI agents such as Claude Code, Codex, OpenCode, and Amp from a single interface.
- Academic research skills
A skill set for Claude Code: research → writing → review → refinement → finalization.
- HERTZBEAT: AI MONITORING IN REAL TIME
Hertzbeat — an AI platform for real‑time monitoring and alerts, unifying metrics, logs, and tracing.
- GLM-5: FROM VIBE CODE TO AGENTIC ENGINEERING
GLM-5 helps transition from vibe coding to agentic engineering. It provides long‑term planning, code generation and self‑correcting agents. Enables building AI systems for complex workflows without constant supervision.
- IPOLLOWORK: Local AI Workspace
iPolloWork combines a local AI agent with design and video tools, delivering a full workflow without servers.
- COMBINE LLM MODELS EASILY
Mergekit offers tools to merge pretrained LLMs, combining weights, averages, and layers from different architectures like Llama. Output is a single model preserving the strengths of the sources.
- Tiny Pixel Effect Snippets
Collection of 244 pixel art effects as minimal code snippets, plus a Claude Code skill for creating custom effects.
- AGENT-ORCHESTRATOR - Autonomous AI Agent Orchestrator
Automatically plans tasks, launches parallel AI agents, and independently fixes CI, merge conflicts, and conducts code review.
- AI Short Video Maker
PRINTFILM uses AI to generate short marketing videos and drama scripts from text prompts, offering automated video assembly, branding, and export options.
- MARKDOWN TASK MANAGER FOR AI AGENTS
Backlog.md is a markdown-native task manager and Kanban board for Git repositories. It structures AI agent work through three review checkpoints — spec, plan, and code — so humans can control large-scale code generation.
- Modular LLM Apps in Rust
⚙️🦀 Build modular and scalable LLM applications in Rust
- Self‑Evolving AI Engine
A GEP‑based engine for self‑learning AI agents with auditable evolution via genes, capsules, and events.
- GIT & GITHUB GUIDE
A practical guide to Git and GitHub for modern engineering teams and AI development. 7,396 ⭐, Python.
- TURBO-FIELDFARE - GEMMA 4 ON MACBOOK
Run Gemma 4 26B‑A4B on any MacBook with an M‑series chip using under 2 GB of RAM. Fast and with no extra resources.
- AI Prompt Master
A skill for Claude that creates precise prompts for any AI tool without losing tokens or credits.
- MCP FOR AI BROWSER CONTROL
MCP server giving AI full browser control: navigation, clicks, text input, screenshots. Works with Claude, Cursor, and other clients via standard protocol.
- Plane and Satellite Tracker
Open platform for monitoring private aircraft, satellite data, and seismic activity with real-time AI analytics support.
- AI partner for 300+ models
ForgeCode is an AI coding assistant that works with Claude, GPT, O Series, Grok, Deepseek, Gemini, and over 300 models.
- DeepSeek Balance Whale Widget
A cute whale widget for DeepSeek Harness that monitors your account balance with drag-and-snap, flip, and rolling number animations.
- Browser Log Monitoring
View browser logs directly in Cursor and other IDEs supporting the MCP protocol.
- MISTRAL.RS — FAST LLM ON RUST
Fast and flexible engine for inference of large language models. 7.2K stars on GitHub.
- COUCOU — CLAUDE CODE IN NOTCH
A tiny companion lives in Dynamic Island on macOS or at the top of your screen on Windows. Tracks Claude Code sessions and signals status without extra windows.
- Terminal Video Downloader
Download videos from various sites directly in your terminal without ads or tracking.
- AI Agent Desktop Interface
ClawX offers a desktop GUI for OpenClaw AI agents, letting users launch, monitor, and manage workflows without touching a terminal.
- MULTI-MODEL AI WORKSPACE FOR EXPERTS
Open-source AI workspace for experts with 800+ model support. Allows switching models mid-chat, tuning 40+ parameters, and running Beam comparison using your own API keys.
- OPENLLMETRY LLM MONITORING
Open-source tool for GenAI and LLM application observability based on OpenTelemetry. 7.1k stars!
- ds4 — Metal inference in C
Local inference engine for Metal in C. Fast startup of DeepSeek-V3-0324 Flash on Apple devices.
- OPENSQUILLA: EFFECTIVE AI AGENT
OpenSquilla — a token‑efficient AI agent that, with the same budget, provides greater intelligence density. Implemented in Python and compatible with popular LLMs for building agentic applications.
- Cross-Platform Native UI Framework
Rune builds native desktop applications using platform UI APIs directly — no webviews, no Electron overhead.
- COURT AGENT SKILL
Trust layer for agent trading: natural‑language mandates, ERC‑7710 permissions, x402 payments, escrow and dispute resolution as an open Agent Skill / Claude Code plugin.
- Speech Tools for MLX
A text-to-speech (TTS), speech-to-text (STT) and speech-to-speech (STS) library built on Apple's MLX framework, providing efficient speech analysis on Apple Silicon.
- OPEN SOURCE BROWSER FOR AI AGENTS
Steel Browser is a ready-made sandbox for web automation. Forget infrastructure and focus on tasks.
- CURSOR-TALK-TO-FIGMA-MCP AI WORKS WITH FIGMA
MCP integration between an AI agent (Cursor, Claude Code, Codex) and Figma, allowing to read and modify designs programmatically.
- MLX-Powered Laya Decisions
A native MLX runtime for Laya typed decision models that runs short decisions in 7–14 ms on an M3 Max. It avoids text generation, PyTorch, and cloud APIs.
- SMOLVM: VIRTUAL SANDBOX FOR AGENTS
Smolvm — embeddable, portable, and branchable virtual machine in Rust for safely running AI agents locally. The project already has 6.5k stars and supports integration with crun/libkrun.
- AI INFOGRAPHIC — ANIMATE WORDS
Infographic — a TypeScript framework for creating infographics using AI. Turn text into visuals in seconds.
- OPENFILM AI VIDEO AGENT
Openfilm turns any coding agents (Claude Code, Codex) into video agents, generating HTML animations and motion graphics. The built‑in editor lets you instantly modify any element of the generated video. The project is written in TypeScript and already has 214 stars on GitHub.
- AI COWORKERS WITH MEMORY & PC
Suparagent — open-source alternative to Grok Bot and Claude Duet. Run named agents with instructions, memory, routines, and isolated computer. They act only after your approval. 6.9k stars on GitHub.
- FLYTE: AI Workflow Orchestration
Flyte is a dynamic and resilient platform for AI orchestration, enabling easy unification of data, models, and compute resources into cohesive workflows.
- Autonomous Red Teaming Platform
T3MP3ST is an autonomous red-team platform that brings multiple agents together into a meta-node to conduct offensive operations and security testing.
- GEO‑SLEUTH DETERMINE PHOTO LOCATION
The agent locates a photo’s shooting spot using OpenStreetMap geometry, elevation profile, satellite imagery and Street View. It works with Claude Code, Codex, Cursor, Gemini CLI, OpenCode and GitHub Copilot, showing each step.
- LLM evaluation platform
OpenCompass is an LLM evaluation platform supporting a wide range of models (Llama3, Mistral, InternLM2, GPT-4, LLaMa2, Qwen, GLM, Claude, etc.) across 100+ datasets.