convez.ai
MCP ServerDesenvolvimento WebAvançado
patil-shubham-dev

AgenticOS

patil-shubham-dev/AgenticOS

SO de desenvolvimento com IA multi-agente, MCP e workspace interativo

Ver no GitHub
InicialPluginScore · de 0 a 100

Um sistema operacional de desenvolvimento baseado em Tauri 2.x e React 19 que transforma modelos de IA em uma força de trabalho de agentes autônomos para desenvolvimento de software. Oferece orquestração multi-agente, suporte completo a MCP, acesso nativo ao filesystem e um ecossistema extensível de ferramentas para automação de tarefas de codificação.

Instalação

$npx convez-ai patil-shubham-dev/AgenticOS

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
25
Forks
4
Acompanhando
2
Saúde
28%
Último commit
há 5 dias
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Automação de fluxos de desenvolvimento e codificação com múltiplos agentes autônomos
  • Integração e orquestração de modelos de IA (Claude, GPT-4o, Ollama) em ambiente desktop
  • Extensão de capacidades de IA através de Model Context Protocol servers

Veredito

Projeto ambicioso e bem-estruturado que leva a orquestração de agentes IA para o desktop com TypeScript/Tauri. Ideal para desenvolvedores que buscam automatizar fluxos complexos, mas requer conhecimento técnico sólido em Rust, Node.js e arquitetura de sistemas distribuídos.

README

A tradução para o português deste README ainda está na fila.

AgenticOS

An open-source, AI-powered development operating system with a rich desktop interface, multi-agent orchestration, MCP integration, and an extensible tool ecosystem.

AgenticOS is a Tauri-based desktop application that turns AI models (Claude, GPT-4o, etc.) into a collaborative agent workforce for software development. It features an interactive file explorer, multi-panel code workspace, chat-based AI assistant, real-time agent orchestration, and granular permission controls.

AgenticOS Tauri React TypeScript


Features

  • Workspace File Explorer — Browse, open, create, rename, and delete files/folders with a VS Code-style tree view
  • Multi-Panel Code Workspace — Split-view code editor with Monaco, diff viewer, and file tabs
  • AI Chat Assistant — Conversational interface with streaming responses, markdown rendering, and syntax-highlighted code blocks
  • Multi-Agent Orchestration — Autonomous, fastest, most-accurate, research-heavy, human-guided, and safe execution modes
  • Agent Workforce — Coordinator, coder, researcher, reviewer agents that collaborate on tasks
  • Extensible Tool System — Built-in tools: Read, Write, Edit, Glob, Grep, Bash, WebFetch, WebSearch, and more
  • MCP Support — Connect to Model Context Protocol servers (stdio, SSE) for expanded tooling
  • Runtime Execution Engine — Real-time agent execution with telemetry, token tracking, and streaming
  • Provider Gateway — Connect OpenAI, Anthropic, Ollama, OpenRouter, and any OpenAI-compatible API
  • Role-Based Configuration — Assign models and providers to specific agent roles (manager, coder, researcher, etc.)
  • Execution Dock — Monitor agent activity, execution depth, tool usage, and performance metrics
  • Snapshot Browser — View and restore file snapshots from agent sessions
  • Built-in Browser & Design Tools — Preview web UIs and design artifacts within the workspace
  • Git Integration — Basic git operations from within the workspace
  • Shell Integration — Register/unregister "Open with AgenticOS" right-click context menu in Windows Explorer
  • Auto-Updater — Built-in update mechanism with install dialogs

Getting Started

Prerequisites

Installation

# Clone the repository
git clone https://github.com/patil-shubham-dev/Agentic-OS.git
cd AgenticOS

# Install dependencies
npm install

Development

npm run dev

Starts the Vite dev server at http://localhost:5173.

Desktop Development

npm run tauri:dev

Launches the Tauri desktop application in development mode with hot-reload.

Build

# Web build
npm run build

# Desktop build (produces .exe + installer)
npm run tauri:build

Outputs:

  • src-tauri/target/release/agenticos.exe — portable executable
  • src-tauri/target/release/bundle/nsis/AgenticOS_2.1.0_x64-setup.exe — NSIS installer
  • src-tauri/target/release/bundle/msi/AgenticOS_2.1.0_x64_en-US.msi — MSI installer

Other Commands

CommandDescription
npm run typecheckRun TypeScript type checking
npm run lintRun ESLint
npm run testRun Vitest tests
npm run previewServe production build locally

Architecture

AgenticOS/
├── src/                    # React/TypeScript frontend
│   ├── agents/             # Agent definitions & worktree manager
│   ├── components/         # UI components
│   │   ├── layout/         # Navigation rail, app shell
│   │   ├── runtime/        # Execution explorer, status bar, dock
│   │   ├── settings/       # Provider, role, and tool settings
│   │   ├── ui/             # Design system primitives
│   │   └── workspace/      # File tree, code workspace, chat, browser
│   ├── core/               # Kernel, routing, providers, error boundaries
│   ├── editor/             # Monaco editor runtime integration
│   ├── hooks/              # Shared React hooks
│   ├── lib/                # Core libraries (workspace, git, provider-gateway, etc.)
│   ├── pages/              # Route pages (code-canvas, settings, onboarding, etc.)
│   ├── performance/        # Performance monitoring, leak detection
│   ├── providers/          # AI provider adapters (OpenAI, Anthropic, Ollama, etc.)
│   ├── runtime/            # Execution engine, agent orchestration, sessions
│   ├── stores/             # Zustand state stores
│   └── types/              # TypeScript type definitions
├── src-tauri/              # Rust backend (Tauri)
│   └── src/
│       └── lib.rs          # Tauri commands, plugins, menu, tray
├── packages/               # Shared packages
│   ├── providers/          # AI provider gateway (shared)
│   ├── shared/             # Common types, constants, utilities
│   └── ui/                 # UI component library
└── scripts/                # Build and utility scripts

Key Architecture Decisions

  • Desktop-first: Built on Tauri v2 for native performance and small binary size
  • Plugin-based FS: Uses tauri-plugin-fs for file operations with configurable scope
  • Multi-agent architecture: Coordinator agent delegates to specialized sub-agents
  • Provider-agnostic: Gateway pattern supports any OpenAI-compatible API
  • Real-time streaming: EventBus-driven architecture for live execution telemetry
  • State management: Zustand stores with selector-based subscriptions

Configuration

Providers

Connect AI providers in Settings → Providers:

  • OpenAI (GPT-4o, GPT-4o-mini)
  • Anthropic (Claude 3.5 Sonnet, Claude 3 Opus)
  • Ollama (local models)
  • OpenRouter
  • Any OpenAI-compatible API

Roles

Assign models to agent roles in Settings → Roles:

  • Manager — Orchestrates and delegates tasks
  • Coder — Writes and edits code
  • Researcher — Searches codebase and web
  • Vision — Analyzes images/screenshots
  • Design — Creates UI designs
  • QA — Tests and verifies
  • Browser — Web automation
  • Memory — Context management

Execution Modes

ModeDescription
AutonomousAI auto-selects agents and tools
FastestOptimize for speed — parallel execution
Most AccurateMulti-agent verification & review
ResearchDeep analysis, extensive searching
Human GuidedApprove every action before execution
Safe ModeRead-only analysis, no mutations

Tech Stack

LayerTechnology
Desktop FrameworkTauri 2.x (Rust backend)
UI FrameworkReact 19, TypeScript 5.6
Build ToolVite 6
StylingTailwind CSS 4
AnimationFramer Motion
Code EditorMonaco Editor
State ManagementZustand 5
ValidationZod
Panel Layoutreact-resizable-panels
Routingreact-router-dom 7
Virtualization@tanstack/react-virtual
MCP@modelcontextprotocol/sdk
IconsLucide React
TestingVitest

Project Status

Active development. The codebase has been restructured from a monorepo architecture into a streamlined Tauri desktop application with a rich multi-agent orchestration system.


License

MIT

Ver o original no GitHub

Bom para

Relacionados