Explore the Platform
OpenShopFloor connects AI agents to a real factory through MCP servers, a Knowledge Graph, and a #shared.UnifiedNameSpace. Explore how each layer works.
What is MCP?
The Model Context Protocol is an open standard for connecting AI models to external data sources and tools. Instead of hardcoding API calls, MCP provides a unified JSON-RPC interface that any AI agent can discover and invoke dynamically. OpenShopFloor exposes 8 MCP servers with 135+ tools covering every domain of a manufacturing operation.
8 Logical Servers
Factory
86 tools · :8020ERP, OEE, capacity, MRP, maintenance, energy, stock, purchasing, subcontracting, TMS, SGM, assembly, pre-assembly, test field.
KG
12 tools · :8035Knowledge Graph: impact analysis, paths, semantic search, Cypher queries, delivery snapshots, chart generation.
Historian
6 tools · :8030Time-series analytics: trends, comparisons, aggregations (AVG/MIN/MAX), anomaly detection, machine/variable discovery.
Discovery
2 tools · AGEAuto-discovered machines and sensors from MQTT UNS via Apache AGE graph.
#shared.UNS
8 tools · MQTT#shared.UnifiedNameSpace: a multi-tenant MQTT backbone where all users and AI agents share the same live factory data stream.
KG
6 tools · AGEKnowledge Graph via Apache AGE: impact analysis, dependency tracing, bottleneck detection, and shortest-path queries.
Historian
6 tools · :8030Time-series analytics — trends, anomalies, aggregations. Industrial MQTT data stored in TimescaleDB with automatic downsampling and retention policies.
KG Builder
8 tools · :8035Knowledge Graph queries — impact analysis, semantic search, shortest path, subgraph extraction. Schema-driven graph built from PostgreSQL sources.
Architecture
Example Tool Call
// JSON-RPC request to MCP server
{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "factory_get_capacity_overview",
"arguments": {
"periodType": "week",
"periodCount": 2
}
},
"id": 1
}