OpenCore MES Try the demo
The problem

Changes reach the plant untested, unapproved and untraced.

A hotfix on Friday, a field edited in the database, an auditor asking who and why. OpenCore MES answers each.

Solution

One path for every change.

A field, a rule, a screen, an approver: all go the same way, on the record.

  1. Request

    Anyone asks from the page itself: a picture, drawn on; their line approves. A problem found at verification comes back here too.

  2. Design

    Drafted alone, with co-designers or the AI, with why.

  3. Review

    Someone who did not design it checks the frozen draft.

  4. Approval

    Each department it touches signs, never its authors.

  5. Execution

    Planned by its named executor, outside the plant's blackout or by a signed exception; live when they press Execute: exactly as approved, everywhere, no restart.

  6. Verification

    They check it in the plant, then sign that it works, or report a problem with a picture of the page.

In the demo: Changes waiting for approval

Changes, as they moveA pipeline for the plant, not for code: every gate a person, every step on the record.
The CI/CD loop, as a plant runs it: a request, design, fitness test and review on one side; approval, execution and verification on the other; what verification finds comes back as a request CHANGE PLANT Request Design Fitness test Review Approve Execute Verify
CI/CD, for the plant

Software runs this loop for every commit: build, test, deploy, and watch. OpenCore MES runs it for every change to the plant's designs, and it can start where software's starts with an issue: someone on the floor asks, from the page they are on, with a picture of it. The fitness test is the automated part, as a build and its tests are: rule cases, scenarios in a sandbox, records replayed. The difference is the gates: a reviewer, each department the change touches, then the person named to execute it when the plant is ready, and the same person checking it works.

Verification closes the loop and opens the next one: a problem found in the plant comes back as a request, with a picture of the page, and goes to design again, as a fix or as a rollback one approval puts live.

Requested

  1. Lot · hold on contaminationQuinn asked, with a pictureline to approve

Design

  1. Screen · line boardDana · fitness test1 scenario fails

Review

  1. ERP · confirm releasesIris · frozen, hash 41abVera reviews

Approval

    Approved

    1. Lot · hold on contaminationplanned Fri 18:30, in the blackoutexception: Sam

    Executed

    1. Machine · stop causeEli pressed Executeto verify

    Verified

      1. Lot · hold on contaminationDana planned it for Friday 18:30, in the plant's blackout: its exception is asked of Sam.
      2. Machine · stop causeEli pressed Execute: checked against what is live, version 7 live everywhere.
      3. Lot · hold on contaminationSam approved it for Production, signed.
      4. Screen · line boardFitness test: the scenario “a held lot” fails: the board hides it.
      Solution

      No hotfix on Friday, unless someone signs for it.

      Set when nothing goes live. Every approved change is planned; one planned inside the blackout is an exception, with a reason and a signature.

      • Your blackoutWhole days and each day's hours, on the plant's clock; set through an approved change.
      • Planned before it goes liveIts executor sets the moment; Execute opens then, not before.
      • Exceptions signedThe executor's reason, approved by someone who neither designed nor executes it; for that blackout only.
      • Emergencies still goTheir reason kept, and reviewed afterwards.
      Solution

      Every version kept. Roll back in one step.

      Nothing is overwritten, nothing is deleted.

      • See what changesPublished beside proposed, before anyone signs.
      • Test togetherDrafts tried side by side in a test sandbox.
      • Undo with one approvalThe rollback is drafted for you, and recorded too.
      Solution

      Changed in the database? It is found.

      Turn tamper evidence on, and the audit trail proves itself and every record proves it still holds what the platform wrote.

      • Every record sealedYour choice, made through an approved change; a write around the platform cannot match the seal.
      • Caught as it happensWho, from where, each field before and after.
      • Scanned all the timeEvery fifteen minutes; every record each week.
      • Closed with a signed reportNever quietly restored; into your own NCRs.
      Solution

      Made for plants that answer to auditors.

      Part 11 practices built in; SOC 2, ISO 27001 and HIPAA mapped control by control in the repository.

      • Your sign-in. Single sign-on, LDAP or Active Directory, or passwords with a second factor.
      • An audit trail that checks itself. Append only; chained and verified every 15 minutes once you turn tamper evidence on.
      • Segregation of duties. Nobody approves what they designed or reviewed.
      • Signed by two. An operator and a verifier, each proving it is them.
      • Record changes that wait. Edits your design names go to the people responsible.
      • Approval flows. Steps in order, each signed by whom the record names: one person, or each of several.
      • Messages you design. What approvers are sent, written and approved like any change.
      • Executed when you are ready, then verified. An approved change waits for the person named to put it live; they sign that it works, or report a problem, reminded each day until they do.
      • Change blackouts. Days and hours when nothing goes live; a change planned inside one needs a reason and a signed exception.
      • Emergency changes. Approved on one signature, reviewed within days.
      • Sensitive fields. Hidden by default, shown with a reason, never sent to the AI.
      • Retention and erasure. Periods approved like any change; the audit trail kept.
      • Backups you have restored. Nightly, encrypted, a restore tested every month.
      • "Why can't I?" Every refusal explained from the rule that decided it.
      A lot held by an out-of-control plan: where it is on its route, the plan and its step, and beside each field why it is read only
      A lot on hold: its route, its plan, and why each field is read only.
      A lot held by an out-of-control planA lot on hold: its route, its plan, and why each field is read only.