REVISION REVIEW
See what changed.
Two revisions. One clear picture. Compare your files or explore a sample below.
Or start with an example
Public test models · precomputed results, with a live recalculation option.
Compare two revisions
- Choose the old revision
- Choose the new revision
- Select Compare
- Read the result and download the inspection list
Choose two files to see what changed and what needs another inspection.
Optional settings
Mesh comparisons use 5,000 area samples per direction by default (20,000 optional), with triangle-surface health and properties; Boolean and semantic PMI stages are not run for mesh inputs. For single-solid BRep inputs, automatic targets 16 area samples per face up to a 20,000-sample budget. Large models use at least 4 per face within the 20,000-face limit; density is recorded per face. Higher counts are not certified accuracy. This choice applies to new distance stages; retained measurements keep their recorded budget.
A saved five-stage session can add this stage through Continue missing stages. Already recorded Boolean evidence is retained when continuing, regardless of this selection.
Existing session stages are retained. PMI attributes are observations, not semantic GD&T validation. Older application packages cannot open sessions extended with PMI.
Multi-solid automatic sampling: 1,000 area samples per body and direction, raised to at least four per face. Body sessions start a fresh run when recalculated; per-body budgets and unresolved pairs are recorded.
Continue or open a saved comparison
To refine a saved revision, select the same two STEP files and a new area-sampling budget, then recalculate surface distances only. Other completed stages are retained. If the new distance run fails or is cancelled, the previous session stays available. Save the new session to keep it.
Choose your old and new revision, then select Compare.
Results available during this run
Some results are ready. Unfinished stages are marked not completed. The previously opened result below remains unchanged until the run finishes. When continuing a saved session, its review notes are preserved. Notes whose evidence has changed require revalidation; saving does not approve them.
Recorded detail
The revision and its notes remain unchanged. Close this panel to return to the summary.
4 · Your result
Observed findings
Where is the difference?
Baseline OCCT tessellation provides visual context only; measurements do not use the mesh. Colored points are measured samples. Drag or use arrow keys to rotate. Select a region to focus. Gaps are unmeasured.
Technical details, health and sampling
Supporting observations
Unresolved scope
Validation report
Compare recorded geometry observations with your tolerance profile. Surface PASS covers measured sample points only, not a certified whole-surface limit.
Names are self-declared. PDF includes a human signature space; it is not digitally signed.
Not evaluated
| Body | Metric | Baseline / declared | Candidate / calculated | Difference | Limit | Result | Reason |
|---|
Exact observations and witness faces are retained in JSON and PDF. Hover over a row for its full numeric record. GVP declarations are optional; missing declarations do not establish a successful exporter comparison. Unit, transfer and measurement gaps prevent a total PASS.
Same ordered STEP bytes required. Opens recorded observations without recalculating. This comparison is separate from the current session report.
Save both session JSON files to reopen this comparison later. The current session JSON contains only its own measurements.
Record an engineering review note
Annotations are separate from measurement findings. Save the session JSON after adding notes. If the referenced evidence changes when continuing, its notes need revalidation.
Names and timestamps are self-declared, not authenticated signatures. Up to 200 notes. A session with notes uses a review wrapper that older releases do not open.
Stage evidence and detailed tools
Scope and interpretation
Health, candidate structure pairing and the selected area-sampling budget per direction, when eligible. Imported coordinates; no automatic alignment. Positive distances and placement differences are recorded observations. No observed difference does not prove unchanged geometry. This workflow does not decide conformity or reprogramming. Boolean volumes are included when selected and both health prerequisites hold; zero or equal volumes do not establish equivalence. Explicit semantic PMI observations and their source-linked display are included; semantic GD&T conformity is not evaluated.
Completed stage records stay available after later failures or cancellation. Continue missing stages requires the same two STEP byte snapshots in the same order; it preserves existing completed evidence without recalculation. Start a new comparison to rerun all stages. Retained and new stages may have different runtimes; each retains its evidence. Open recorded detail transfers a completed health, structure or distance record to its tool without recalculation. Keep the revision tab open until it is received. JSON download remains available for every completed stage. The desktop edition processes files on this computer; the web edition processes them in your private server session. Session and component records may contain confidential source names.