Why Qontiv
One system — not a QMS bolted to an MES.
Qontiv is a manufacturing execution system for regulated discrete manufacturers — automotive, aerospace, and medical device — with quality built into the same platform that runs production rather than sold alongside it.
That matters for a specific reason. Your quality system can hold a record saying a gauge was due for calibration in March. It cannot tell you whether the press that ran Tuesday's lot was using that gauge, because it is not the system that ran the press. Qontiv is. Everything below is us showing our work rather than asserting it.
The gap a second system cannot close
A standalone quality management system is a system of record. It stores what someone told it: the calibration was done, the operator was trained, the change order was approved. Every one of those facts arrives as an assertion typed in after the event, usually by the person whose job it was to do the thing.
That is not a criticism of QMS software. It is a structural limit. A system that does not run production cannot observe production. When your auditor asks whether the control was operating — not whether it was documented — a system of record can only hand back the documentation.
An MES sits in the other position. It is the thing dispatching the work order, holding the operator's sign-off, taking the reading off the gauge, and stamping the equipment that made the part. It does not have to be told whether an uncertified operator ran a certified step. It was there.
So the case for Qontiv is not that it has more quality features than a QMS. It is that a handful of questions your auditors care about can only be answered from where production actually happens — and if quality lives in a separate system, nobody is standing there.
6 controls Qontiv computes from production data
These are not questionnaire answers or checkbox attestations. Each one is evaluated against the rows your plant generated, over a rolling observation window, and drills through to the records behind it so a reviewer can follow the number back to the source.
We will not print results here, and it is worth saying why: computed status isper-tenant and in-product only. It is your data, it lives behind your login, and a figure on a marketing page could only ever be ours or invented. Ask for it in a demo against a seeded dataset, or run it on your own during a pilot.
Did the audit trail ever go quiet?
Calendar days in the observation window, checked for stretches with no audit activity at all.
21 CFR Part 11 §11.10(e)
Are privileged accounts still being used?
Active users holding planner-level roles or above, checked for a successful sign-in — including those that have never signed in.
21 CFR Part 11 §11.10(g)
Did any approved revision get promoted without a signed change order?
Items, BOMs, and routes that are approved and cite a change order, checked for an approval that is actually signed and not a rejection.
IATF 16949 §8.5.6.1
Was every gauge calibrated on schedule?
Every gauge master that is not retired, checked for a calibration date that has passed — or was never set.
ISO 13485:2016 §7.5.1
Are layered process audits happening when they were scheduled?
Audits scheduled inside the observation window, checked for overdue, late, or completed-without-a-completion-date.
ISO 13485:2016 §8.2.3
Did an uncertified operator run a step that required certification?
Work-instruction executions whose instruction declares a skill requirement, checked against the operator who actually ran it.
ISO 13485:2016 §6.2
What the limits are
6 controls is 6 controls. The rest of the catalog is mapped and evidenced but not computed, and we are not going to describe 6 as coverage. The computed set reaches exactly the 21 CFR Part 11, IATF 16949 and ISO 13485:2016 clauses named above — not those frameworks generally, and nothing in AS9100. Two of the evaluators mention equivalent clauses in other standards in their source comments; those are not catalog bindings and should not be read as computed.
A control can also come back not measured — no rows existed to evaluate, or the history was too short to support a defensible answer. That is deliberately not styled as a failure, because it is not one. It means we do not know, and a tool that quietly scores "unknown" as "pass" is worse than no tool.
The rest of what you can check
A control catalog you can read before you buy
125 controls mapped across 10 frameworks, each one naming the Qontiv feature that implements it, the clause it answers to, and who operates it — us, you, or both. The Trust Center table is generated from that catalog at build time, so the page cannot drift from the data. The full catalog is available under NDA.
Documentation and architecture-test references are CI-verified to resolve, and architecture-test references are additionally verified to contain a live test rather than a file that once had one. References naming a GraphQL query are checked against the committed schema. What is not checked is the remainder — references pointing out to in-app routes and API paths — so this is not a claim that every link in the catalog is verified.
See the framework table →A lint that fails our build if we claim a certification we do not hold
Every pull request against this website runs a deny-list against both the source content and the rendered pages. The rules block any sentence asserting that Qontiv itself is certified or conformant, any construction pairing a standard's name with a claim of readiness for it, any assertion that our e-signature binding has been validated by us rather than by the customer, and a specific set of competitor and pricing claims we decided we would not make. It is not a style guide. It blocks the merge.
Its limits, because a control we oversold would undercut the point of this page: there are documented escape hatches — a small allowlist for internal governance files, which have to name a forbidden phrase in order to forbid it, a negation rule so a sentence prohibiting a claim is not itself flagged, and a per-line exemption comment. The content half does not scan page templates; the rendered half covers those, by loading the built pages.
Uptime you read off a status page, not off a slide
Checks run on a schedule, the history is published, and the page names the exact hostname it checks for each component. It also states plainly that those checks currently observe our non-production tier and that production availability is not reported there yet — because publishing a number without saying what it measured is how uptime claims become worthless.
status.qontiv.com →We retract claims when we find them false, and we say so on the page
This is a practice rather than a metric, so there is no score to quote. What it looks like in public: our medical-device page says the Veeva Vault connector is not available yet, because checking the wiring showed it was never reachable even though our own internal documentation described it as working. Our aerospace page carries a section titled "what is not on this page, and why" that names first article inspection and counterfeit-part prevention as things we have not built. Both of those cost us a capability bullet.
We would rather you find the gap here than in week three of a pilot. If you are evaluating us against a vendor whose feature matrix has no gaps in it at all, that is worth a question.
And what we do not have
No customer logos, no testimonials, no case studies, and no ROI calculator — because we have no production customers yet, and every one of those would have to be invented. The scenarios on our customers page are labeled as illustrative for the same reason. Qontiv also holds no SOC 2 Type II, ISO/IEC 27001:2022, or ISO/IEC 42001:2023 certification; those are on the roadmap and the Trust Center says so without dressing it up. A calculator seeded with numbers we made up would be the easiest thing on this list to build and the least worth reading.
How to check any of this
There is no free trial and no self-signup, and that is a deliberate choice rather than a missing feature — an MES configured badly by someone guessing at your routings tells you nothing about the MES.
Evaluation is a paid four-week pilot on one of your production lines, using your real equipment, routing, and quality data, with guided data modeling from our team and go/no-go criteria agreed before it starts. The fee credits in full against a first annual contract. The computed controls above are the natural thing to point at the pilot line: they either find something on your floor or they do not, and either answer is information.
Point it at one line and see what it finds.
A paid four-week pilot on your real routings and quality data, with go/no-go criteria agreed before it starts and the fee credited in full.