Op by Op

Process knowledge platform · for medtech manufacturing engineers

Your PFMEA is the least valuable thing you'll build.

Op by Op captures the engineering reasoning behind every operation — what's critical, what can fail, why each control exists — as a living knowledge base that gets deeper as your line matures. The PFMEA just falls out the other side.

Book a demo See the reasoning chain

Private by design · your data is never used to train AI models · your PDF is discarded after analysis · built for ISO 13485.

An Op by Op analysis on screen — every operation with its CTQs, failure modes, hazards, and mitigations, risk-banded and version-controlled.
The actual analysis — every operation, its CTQs, failure modes, hazards, and the mitigations that control them, risk-banded and fully version-controlled.

Today it's Excel and tribal knowledge.

When engineers leave, the knowledge leaves. When the line changes, the PFMEA goes stale the next day. And when an auditor asks why a control exists, no one remembers.

That's the tell: a PFMEA records that you rated a failure mode — and throws away the part worth keeping, the why. Why that control, why that severity, what you considered and ruled out. Op by Op captures that reasoning at every operation and keeps it version-controlled as your line evolves. The spreadsheet is just the export; the knowledge is what makes your next line change, audit, and new-hire ramp faster.

“I'm a manufacturing engineer. I built Op by Op after watching team after team lose hard-won process knowledge the moment someone left — and watch their PFMEAs go stale the day after a line change. So I built the tool I wanted: walk the line one operation at a time, capture the reasoning at every step, and let the PFMEA fall out the back. The knowledge base you build is worth far more than any single PFMEA.”

Tyler Taverne · manufacturing engineer, founder

Operation by Operation Analysis, done in minutes.

At every operation we walk the full reasoning chain — what's critical, what could fail, the downstream hazardous situation, who's harmed, the mitigations, and the engineering rationale at every link. Then it stays a living, version-controlled record.

  1. 1Upload the procedure. Your MPI PDF is used to generate the analysis, then discarded — only the result is saved.
  2. 2AI drafts the analysis. It lays out the full operation list — steps, CTQs, failure modes, hazards, mitigations, and the rationale for each — in minutes, so your time goes to engineering judgment, not hand-copying steps into a spreadsheet.
  3. 3You refine as the expert. Edit anything; every change is tracked. Re-upload a changed procedure and it redlines exactly what moved.

Anatomy of an operation

The reasoning is the artifact. Here's the shape of one operation.

A worked example — the structure Op by Op captures at every operation. Click any link in the chain to see the rationale for why it was selected, and what was considered and ruled out. (The screenshot above is a real analysis; this is an illustrative walkthrough.)

OP-027 · Line 03 / Catheter assembly · Rev. C
UV cure of catheter hub bond
CTQs identified 3 mitigations Reviewed 04-12

Rationale · why this

Cure intensity at the bond interface, not lamp output.
Bond strength is a step function of cross-link density, which is driven by the photons actually reaching the bond line — not the lamp's nameplate output. We control the variable that drives the outcome.
Below the cure-intensity threshold at the bond line, pull strength drifts out of the acceptance band within weeks of shelf life — so the CTQ is intensity at the interface, not the lamp's nameplate output.

Ruled out · with reasoning

Cure timeRuled out
Correlates with intensity but is downstream of it. Controlling time without intensity allows under-cured bonds at full process time.
Lamp nameplate outputRuled out
Doesn't account for fixture geometry, window fouling, or reflector drift between PMs.
Visual inspection of bondRuled out
Operators cannot reliably detect a 5% cross-link deficit by eye. Detection-based control fails Process Validation Stage 2.

A manual PFMEA is only as complete as whoever wrote it — one engineer breaks the line into sixty operations, another into twelve, and the coarse pass quietly skips the failure mode that surfaces later as a line-down or a recall. Op by Op runs the same operation-by-operation decomposition every time — every engineer, every line — so your coverage doesn't depend on who held the pen.

Built for regulated environments.

Your data is never used to train AI models and is never shared. Your uploaded document is used only to generate the analysis and is then discarded — a failed run is kept only briefly so you can retry, then purged. Only the structured analysis is saved, isolated to your organization. Every change is captured in a tamper-evident, audit-grade log. Read our privacy & data-handling policy and DPA.

Book a demo

We work with each team directly — a short call to map your first line, then you're running. Pricing scales with your team and product lines; we'll cover it on the call.