The problem
A machine stops at 2:14 am. The operator writes "breakdown" on a sheet, or forgets to, or fills it in from memory at the end of the shift. By the time the numbers reach the plant head it is the next morning, and nobody knows how long the machine really stood still, or why.
Most plants still run on counts written by hand, stop reasons guessed after the fact, and reports that arrive when it is too late to act.
What we built
A manufacturing platform that connects to the machines a plant already runs and shows how the plant is performing in real time. It calculates OEE (availability, performance and quality) from the machines themselves, for every machine, line and plant, without spreadsheets or waiting for the end of the shift.
Operators get a panel at the machine with two large buttons, Start and Stop, and stop reasons shown as pictures: breakdown, no material, changeover, break and cleaning. A single tap records the reason at the moment it happens. Work orders come in from the ERP, and every batch is traced to its line, shift and operator.
When a sensor reading crosses a limit, the right person is alerted in the app or by email, SMS or push notification, and the alert escalates if nobody responds. Root cause analysis guides the team through a fishbone diagram and the five whys, then checks that the fix held. An AI copilot answers questions about the plant in plain English, using the plant's own data.
How it was built
An edge agent in the plant reads the machines over OPC UA, Modbus TCP, Siemens S7, EtherNet/IP, MQTT or HTTP and forwards the data securely, with no rewiring and no new controllers. Machine state comes directly from each machine's own run signal.
The backend uses Python and FastAPI, with TimescaleDB for time-series readings, React for the web dashboard and React Native for the mobile app. Nisthora is a configurable product, so a technician can add plants, lines, machines and sensors from the screen, and every role from owner to operator gets its own view.
Why it matters to you
If your plant decisions are based on yesterday's spreadsheet, Nisthora closes that gap. And if you need live data from equipment that was never designed to share it, we have already solved the hard parts.