Skip to content

What is semantic PMI?

PMI is the manufacturing information attached to a 3D CAD model, such as dimensions, GD&T, tolerances, datum features, and other production-related annotations. Semantic PMI is the computer-readable form of that information.

NIST defines semantic PMI as PMI representation that is associated with CAD model geometry and can be consumed automatically by downstream applications. It distinguishes this from graphic PMI, which captures visual appearance but is not meant to be computer-interpretable.

Type What it captures Intended consumer
Semantic PMI The meaning behind annotations — tolerance values, referenced geometry, requirement structure Downstream software (automated use)
Graphic PMI Visual appearance — annotation lines, symbols, positions People (human review)

What does "import" mean in this context?

Here, import does more than display annotations on screen. A semantic PMI import workflow:

  • extracts the PMI entities from the source file,
  • maps them into the target application, and
  • preserves the links between each annotation and the referenced faces, edges, datums, or features of the model.

That association is what lets downstream software interrogate the data instead of treating it as static text or graphics.

Why is that different from graphical PMI import?

Graphical PMI import is mainly about reproducing what a user sees: annotation lines, symbols, positions, and presentation. Semantic PMI import is about preserving the meaning behind those annotations — the actual tolerance value, the referenced geometry, and the logical structure of the requirement. That distinction matters because software can automate tasks from semantic PMI, while graphical PMI is primarily for human review.

Why does semantic PMI import matter in MBD?

Semantic PMI import is closely tied to model-based definition (MBD) workflows, where the 3D model becomes the main design-definition deliverable. In that setting, downstream teams do not just need to view annotations; they need applications to read and use them for tasks such as inspection planning, machining, and quality validation. Autodesk and PTC both describe semantic PMI as enabling this kind of automated downstream use.

Applications and Industry Use Cases

Semantic PMI import is used wherever 3D models need to drive downstream processes without manual re-entry of tolerances and annotations. Typical use cases include:

  • CMM and metrology software
  • inspection planning
  • CAM programming
  • quality control
  • broader model-based enterprise workflows where design intent must survive translation between CAD and downstream systems

It is especially valuable in workflows based on standards such as STEP AP242, where PMI and geometry are exchanged together. NIST's validation work shows that model-centric interoperability depends on both geometry and PMI being transferred accurately enough for downstream manufacturing and inspection systems to consume them directly.

Challenges or Common Pitfalls

Pitfall What to keep in mind
Assuming visible annotations mean semantic data imported correctly NIST notes that each PMI element must be checked for syntactical and semantic accuracy, including correct association to the referenced 3D geometry. Display is not the same as valid downstream interpretation.
Expecting clean mappings during translation NIST's work on STEP-to-ACIS workflows found that even when PMI values were largely preserved, anomalies could occur — association changes, unsupported data structures, or lost unit information.
Assuming universal PMI support Not every file type or CAD implementation exposes the same level of semantic PMI. Autodesk documents file-type-specific support and limitations, so coverage is often format- and system-dependent.

How Spatial Helps

We support this workflow through 3D InterOp. It imports both semantic PMI and graphical PMI, along with Views, Captures, and links to 3D geometry. Once that data is in, your application can query imported PMI values and tolerances directly — for example, to drive automation like measurement planning.

On the format side, we support PMI in STEP AP242 and JT, on top of existing PMI support for CATIA, NX, Creo, and SOLIDWORKS.

Our more recent release notes add another piece: you can now write imported semantic PMI data back out to STEP AP242. That means 3D InterOp is useful not only for reading semantic PMI into an application, but also for carrying that PMI forward into standards-based downstream exchange.

Want to try it on your own models? Request an evaluation or talk to our team.