> ## Documentation Index
> Fetch the complete documentation index at: https://docs.opencompany.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> Visual workflow automation platform with AI agents, WhatsApp, and Android integration

# Welcome to MachinaOs

MachinaOs is an open-source workflow automation platform inspired by n8n, built with React Flow and Python FastAPI. Create powerful automations using a visual node-based editor with real-time execution feedback.

<CardGroup cols={2}>
  <Card title="110+ Nodes" icon="cube">
    AI models, specialized agents, Google Workspace, WhatsApp, Telegram, Twitter/X, Android, documents, webhooks, and more
  </Card>

  <Card title="WebSocket-First" icon="bolt">
    Real-time status updates over a single WebSocket connection
  </Card>

  <Card title="AI-Powered" icon="brain">
    11 LLM providers with conversation memory, tool calling, and thinking modes
  </Card>

  <Card title="Self-Hosted" icon="server">
    Full control with the `machina` CLI, Temporal distributed execution, and encrypted credentials
  </Card>
</CardGroup>

## Key Features

### AI Integration

Connect to 11 LLM providers with built-in conversation memory, tool calling, and thinking modes.

* **OpenAI** - GPT-5.5/5.4/5.2, GPT-4.1/4.1-mini/nano, o3/o4-mini, GPT-4o/4o-mini with response format and reasoning effort
* **Anthropic** - Claude Fable 5, Opus 4.8/4.7, Sonnet 4.6, Haiku 4.5 with extended thinking via `budget_tokens`
* **Google Gemini** - Gemini 3.5 Flash, 3.1 Pro/Flash-Lite, 3 Flash, 2.5 Pro/Flash/Flash-Lite with multimodal support and thinking budget
* **OpenRouter** - 200+ models via unified API (OpenAI, Claude, Llama, Mistral, and more)
* **Groq** - Ultra-fast inference with Llama 3.3-70b, Llama 3.1-8b, GPT-OSS-120b/20b, Qwen3-32b
* **Cerebras** - Ultra-fast inference on custom wafer-scale hardware (GPT-OSS-120b, Qwen-3-235b, GLM-4.7)
* **DeepSeek** - V4 Flash + V4 Pro (up to 1M context)
* **Kimi** - Moonshot K2.6 / K2.5 / K2.7-code (256K context)
* **Mistral** - Large / Medium / Small / Codestral (up to 256K context)
* **Ollama** - Run your locally-pulled models via a local Ollama server
* **LM Studio** - Run models loaded in the local LM Studio server
* **AI Agent** - Agent loop with tool calling and iterative reasoning
* **Chat Agent** - Conversational agent (Zeenie) with skill support for multi-turn chat
* **16 Specialized Agents** - Android, Coding, Web, Task, Social, Travel, Tool, Productivity, Payments, Consumer, Autonomous, Orchestrator, AI Employee, RLM, Claude Code, Codex
* **Simple Memory** - Markdown-based conversation history with long-term vector storage

### WhatsApp Automation

Send and receive WhatsApp messages with event-driven triggers.

* QR code pairing via Go WhatsApp bridge (whatsmeow)
* Message filtering: sender, group, keywords, forwarded status
* Group/sender name persistence for display

### Android Device Control

Automate Android devices via local ADB or remote WebSocket relay.

**16 Service Nodes:**

* System Monitoring: battery, network, system info, location
* App Management: launcher, app list
* Device Automation: WiFi, Bluetooth, audio, device state, screen control, airplane mode
* Sensors: motion detection, environmental sensors
* Media: camera control, media playback

### Webhooks & HTTP

Event-driven webhook triggers and HTTP client for external APIs.

* Dynamic webhook endpoints at `/webhook/{path}`
* HTTP Request node with GET, POST, PUT, DELETE, PATCH
* Webhook Response for custom responses

### Code Execution

Execute Python, JavaScript, and TypeScript directly in workflows.

* Syntax-highlighted editor with per-theme code colors
* Access to `input_data` from connected nodes
* Console output capture
* Sandboxed Python execution via Monty with enforced time and memory limits

## Quick Links

<CardGroup cols={2}>
  <Card title="Installation" icon="download" href="/installation">
    Get MachinaOs running locally in minutes
  </Card>

  <Card title="Quick Start" icon="rocket" href="/quickstart">
    Build your first workflow in 5 minutes
  </Card>

  <Card title="Node Catalog" icon="grid" href="/nodes/overview">
    Explore the full node library
  </Card>

  <Card title="Tutorials" icon="graduation-cap" href="/tutorials/first-workflow">
    Step-by-step guides
  </Card>
</CardGroup>

## Architecture Overview

MachinaOs follows n8n-inspired architectural patterns with modular backend services:

| Layer      | Technology                                                                      |
| ---------- | ------------------------------------------------------------------------------- |
| Frontend   | React 19, TypeScript, React Flow, Tailwind + shadcn/ui, Zustand, TanStack Query |
| Backend    | Python FastAPI, SQLAlchemy, plugin-first node system                            |
| Real-time  | WebSocket-first status + RPC                                                    |
| Execution  | Temporal distributed workflow execution                                         |
| WhatsApp   | Go service with whatsmeow library                                               |
| Cache      | Redis (production) / SQLite (dev) / Memory fallback                             |
| Deployment | `machina deploy up --provider gcp` (Terraform + `machina serve` under systemd)  |

### Backend Service Architecture

```
server/services/
├── workflow.py              # Facade (~460 lines)
├── node_executor.py         # Registry-based dispatch
├── parameter_resolver.py    # Template variable resolution
├── deployment/              # Event-driven lifecycle
│   ├── manager.py           # Deploy, cancel, status
│   └── triggers.py          # Cron and event triggers
└── execution/               # Parallel orchestration
    ├── executor.py          # Conductor decide pattern
    └── cache.py             # Redis/SQLite persistence
```

<Info>
  MachinaOs is designed for self-hosting. Your data stays on your infrastructure with automatic cache fallback for any environment.
</Info>