The challenge
A German glass quality inspection company audits glass manufacturers, and most of that work happens on the factory floor. We built and delivered glass quality inspection software that its teams use to run shifts, check incoming material and record quality tests in one web system.
Insulating glass is checked at many points. Glass and other materials arrive and need inspecting on receipt. Cut glass, processed glass, the finished unit and the gas sealed inside it are tested every day, and each material has its own tolerances.
Spread that across several manufacturers, each with its own suppliers and staff, and paper forms start to strain. A reading outside tolerance has to be caught when it is entered, not when someone leafs through a binder at the end of the week. Any correction also needs a reason that stays on record.
What we built
The system is a web application on Node.js, Angular and MongoDB, running on an Ubuntu cloud server. Master data sits underneath everything: the manufacturers the company audits, their users, suppliers, and material masters that hold each material's tolerances.
Shift management. The company manages its shifts in the same system that holds the inspections, so each day's checks sit alongside the shift that ran them.
Incoming material checks. Two forms cover incoming glass and incoming semi-finished products, filled in as each delivery is received.
In-process tests. Daily tests after the glass is cut and after it is processed.
Finished unit and gas-fill tests. Daily tests on the finished insulating glass and on the degree of gas filling.
Tolerance alerts. A reading outside the material's tolerance fires a trigger. Inspectors can still change the data, but only with a recorded reason.
Save now, submit later. Readings are saved at intervals through the day, and the form is submitted once it is complete.
Roles and reports. Role-based login decides what each person can do, and client users get the rights that fit them. Reports follow the details and filters the company asked for.
The six inspection forms follow the path glass takes through a plant, from the delivery bay to the sealed, gas-filled unit. Keeping that order is what makes the records easy to read back later.
How it works day to day
Take an inspector starting a shift at one of the audited plants. They sign in and see only the forms their role allows.
A glass delivery arrives, so they open the incoming glass form and record it against the right supplier and material. Later they run the cutting test, then the processing test, saving readings as they go instead of keeping them in a notebook.
One value lands outside the material's tolerance, and the system flags it at once. If the inspector corrects the entry, they add a reason, and the reason stays with the record.
Before the shift closes, they finish the tests on the completed unit and its gas fill, then submit. Users with the right access can pull reports filtered by the details they care about.
The result
The software was delivered and is live. Shifts, incoming material and every inspection stage now sit in one system, linked to the materials and tolerances they are checked against.
Out-of-range readings are caught at entry rather than at review, and every change carries its reason. Client users work within the rights they have been given, and reports come from the same records the inspectors entered.
We have no measured before-and-after figures for this project, so we describe what the system makes possible rather than quote a percentage.
What the research says
Cloud is now the norm for European firms. Eurostat reports that 52.7% of EU enterprises paid for cloud computing services in 2025, 7.4 percentage points more than in 2023 (Eurostat, February 2026). Running the inspection system on a cloud server follows that pattern: inspectors at different sites work on the same records.
Access control is where web apps fail most. The OWASP Top 10:2025 keeps broken access control in first place and reports that 100% of the applications tested had some form of it (OWASP Top 10:2025). With the audited manufacturers' own staff logging in too, role-based access decides what each person can see and change.
Where AI fits next
One possibility, not part of this build: a model that reads months of tolerance alerts and correction reasons, then points to the materials, suppliers or stages where problems cluster. Inspectors would still make every call. The AI would only suggest where to look first.
Planning glass quality inspection software?
A few questions to answer before you build:
Which inspection stages do you run, and which happen on receipt rather than daily?
Where do your tolerances live today, and who is allowed to change them?
What should happen when a reading falls out of range, and who has to explain a correction?
Which reports do your auditors, managers and customers ask for most often?
Our web application development team builds systems like this one, and our custom software development work covers the tolerance logic and roles underneath. You can also see more of our work for manufacturing companies.
