Getting Started

OpenShopFloor (OSF) is an open-source factory AI sandbox. It provides a complete manufacturing simulation with real-time machine data, production orders, quality metrics, and tool management — all accessible through 118 MCP (Model Context Protocol) tools. You can build AI agents that monitor, analyze, and optimize factory operations.

Go to openshopfloor.zeroguess.ai/register and create a free account. You'll need an email address and password. After registration you're automatically logged in and can start exploring.

The quickest way to interact with the factory is through the AI Chat. The chat assistant has access to all 118 MCP tools and can answer questions about your factory in real time:

• "What's the current OEE across all machines?"
• "Show me all open production orders"
• "Are there any quality issues today?"
• "Which machines are idle right now?"

The assistant will call the appropriate MCP tools, retrieve live data from the factory simulation, and present the results with analysis.

Agents are pre-built AI routines that perform specific factory tasks. Go to Agents and try one of the built-in agents:

• OEE Optimizer — Analyzes OEE data and suggests improvements
• Quality Guard — Monitors defect rates and identifies quality issues
• Delivery Tracker — Checks at-risk orders and suggests corrective actions

Click "Run" on any agent to see it execute. The output streams in real time as the agent calls MCP tools and analyzes data.

Flows let you chain multiple AI operations in a visual editor. Go to Flows and click "Open Editor" to launch the Node-RED-based flow editor.

1. Drag an osf-mcp-fertigung node onto the canvas
2. Configure it to call factory_get_latest_oee
3. Add an osf-prompt node and connect it
4. Set the prompt to: "Analyze this OEE data: {{input}}"
5. Save and run — watch the LLM analyze your factory data

For more complex logic, write a TypeScript agent and deploy it from GitHub:

1. Connect your GitHub account in Settings
2. Create a repo with osf-agent.yaml and src/main.ts
3. Deploy from Agents → Deploy from GitHub

• Architecture Overview — Understand how the system works
• Node Reference — All 16 custom Node-RED nodes
• API Reference — REST API endpoints
• Factory Simulation — What the simulated factory looks like
• Challenges — Test your skills with factory optimization challenges