About Qmach-AI

A manufacturing intelligence platform, built for mid-market manufacturers.

Qmach-AI is a multi-tenant SaaS platform that helps a plant or quality team find out why a manufacturing loss happened, prove it with evidence, put a rupee figure on it, and confirm afterwards that the fix actually worked.

What we build

Qmach-AI started life as Factory Investigator: an evidence-backed root-cause engine for production, quality and maintenance losses. It has since grown a second half, Manufacturing Operations — the transactional modules a factory runs on day to day. Both halves share one login, one tenant model and one database, so operational data and investigation evidence live in the same connected context instead of two disconnected systems bolted together.

It is deliberately not a full ERP, a payroll or HR system, a CRM, a complete SCADA/PLC control layer, or a 3D digital twin. We would rather do the loss-to-recovery loop well than everything adequately.

How it's built

Factory InvestigatorThe original product: turn a defect spike, breakdown or scrap event into ranked, evidence-backed hypotheses — then quantify the loss and verify the fix actually recovered it.

Manufacturing OperationsThe day-to-day transactional spine a factory runs on: item master, inventory, work orders and procurement — sharing one login, one tenant model, one database with the Investigator.

Tenant isolation by defaultEvery tenant's data is walled off with Postgres row-level security, not application-layer filtering — enforced at the database, not hoped for in code.

One factory contextSupplier, lot, machine, tool, shift and operator are connected in one queryable graph, so a root-cause search doesn't stop at the edge of one module's table.

Where we come from

Qmach-AI is built by the team behind Magnetar — an operator with a background in running and instrumenting real production lines, not just building software for them. That shows up in the product's shape: it is opinionated about the five questions a manufacturing loss always raises — where the money went, what caused it, what evidence supports that, who has to act, and whether the action actually recovered the value — because those are the questions an operator actually asks on the floor.

Where it runs

One codebase, one deployment target: a Next.js web app in front of two FastAPI services and a single shared PostgreSQL database, all tenant-isolated at the row level. The current pilot build runs at qmachai.magnetarai.co.in.