Free Manufacturing AI Playground

Build Manufacturing AI Agents in Minutes. No Factory Required.

A live factory simulation with 8 CNC machines, injection molding, assembly lines, and 111 AI tools. Ask questions, build agents, take challenges — all free.

A Real Factory, Running Live

This isn't a demo dataset. A full manufacturing simulation runs 24/7 with CNC machines, injection molding, quality control, and tool management.

Build AI Agents Your Way

Drag-and-drop in the visual flow editor, or write TypeScript code agents. Wire MCP tools, LLMs, and decision logic together — your choice.

Node-RED Flow Editor with AI agent nodes
Drag & Drop
Connect agents, prompts, and MCP tools visually
Live Execution
Run flows against real factory data, inspect every step
No Code Required
Built on Node-RED — the tool manufacturing already knows

Prefer code? Write TypeScript agents.

Full access to 118 MCP tools and LLM calls from TypeScript. Deploy from GitHub with one click — no Node-RED required.

ctx.mcp.call()ctx.llm.chat()ctx.storageGitHub Deploy
// agent.ts
export default async function main(ctx) {
const oee = await ctx.mcp.call('factory_get_latest_oee');
const analysis = await ctx.llm.chat(`Analyze: ${oee}`);
return { analysis };
}

Deep Analysis Agents

Multi-agent systems that analyze your factory in real-time — powered by 4 MCP servers with 111 factory tools and local LLM inference.

OTD Deep~2-3 min

On-Time Delivery optimization — 4 specialists + moderator discussion + synthesis

OEE Fast~30-60s

Root-cause analysis for low OEE across machines

Shopfloor Fast~30-60s

Morning meeting report — OEE, KPIs, stock levels, open orders

Impact Fast~30-60s

KG bottleneck & cascade analysis — affected orders, customers, alternatives

How the Data Flows

Every machine, sensor, and order generates live data. Six specialized AI servers organize it into 111 queryable tools. Your agents tap into this data and make decisions in real time.

L1
Shop Floor
Machines & Sensors
CNC Machines (9001–9018)
Injection Molding (SGM-001–005)
Assembly & Quality Inspection
Temperature, Pressure, Cycle Times
L2
Data Layer
6 Servers → 111 AI Tools
ERP: Orders, Materials, Scheduling
OEE: Availability, Performance, Quality
Live Data: Real-time machine streams
KG: Dependency & Impact Analysis
L3
AI Agents
Reason, Decide, Act
OEE Monitor & Diagnostics
Material & Capacity Planning
Delivery Risk & Quality Guard
Your Custom Agents
111
MCP Tools
ERP, OEE, QMS, TMS, Live Data, KG
6
MCP Servers
Live production data
24/7
Simulation
Always running

Your AI Playground for Manufacturing

Everything you need to explore, experiment, and build manufacturing AI agents — from first idea to working prototype.

🏭

Real Factory, Real Data

A live manufacturing simulation with CNC machines, injection molding, quality control, and tool management — not a toy dataset.

🔗

118 MCP Tools

Access ERP orders, machine status, OEE metrics, quality reports, tool inventory, live #shared.UNS data, and knowledge graphs through the Model Context Protocol.

🌐

Connected Data Layer

All factory data flows through a unified data layer. Six AI servers organize machines, orders, quality, and inventory into 111 tools your agents can query.

🤖

Build & Experiment

Design your own AI agents — analyze production, predict failures, optimize workflows. Your ideas, your logic, instant feedback.

🎨

Visual Flow Editor

Wire agents, MCP tools, LLMs, and decision nodes together in a Node-RED drag-and-drop editor. No coding required.

💻

TypeScript Agent SDK

Write agents in TypeScript, push to GitHub, auto-deploy. Full SDK with MCP access, LLM, and persistent storage.

🧠

Bring Your Own LLM

Use our platform LLM or connect your own — OpenAI, Anthropic, local models, any OpenAI-compatible endpoint.

🔓

100% Free

No credit card, no trial limits on features. A free playground for manufacturing AI — use it, learn from it, build on it.

How It Works

01

Explore the Factory

Watch real-time production on the live dashboard. Understand the machines, orders, and quality flows your agents will work with.

02

Try an Agent

Run a built-in deep analysis agent or use the chat interface to query factory data with natural language.

03

Build Your Own

Create your own agent — visually in the flow editor or as TypeScript code on GitHub. Connect MCP tools, add LLM logic, wire it up.

04

Experiment & Iterate

Test with live data, inspect every node’s output, tweak your prompts, add decision logic. The sandbox is yours.

05

Share & Discuss

Share your agents with the community, suggest features, report issues, and help shape the future of manufacturing AI.

Why I Built This

I spent years in manufacturing consulting. Every project started the same way: weeks of NDAs, months of waiting for IT to provision access, and by the time you could actually touch real data, the budget was half gone. I was tired of it.

So I built what I wished existed — a real factory simulation that anyone can use to learn manufacturing AI. No NDAs, no waiting for IT. Just open your browser and start building.

The code is public on GitHub. Fork it, learn from it, contribute. But more importantly — just use the playground and start building.

Ready to Explore?

Jump in, try the agents, build your own, and see what manufacturing AI can do. Free to use — no credit card, no strings attached.

This site uses a cookie to remember your preferences. Analytics are anonymous and cookie-free. Privacy Policy