01
Recognize objects
Identify engineered entities and distinguish meaningful occurrences inside assemblies and models.
GEOMETRYLAB • INTERNAL RESEARCH PROGRAM
Internal research. Better tools for engineering work.
GeometryLab is Stratic Systems' internal, research-grade program for building capabilities in CADSync and other systems. It explores how software can recognize engineered objects, measure geometry, and keep the evidence behind a result available for review.
This is a research overview, not a public product or a download offer. Methods and recognition coverage are still being qualified before they can support broader use.
RESEARCH AIMS
A CAD file contains geometry. Understanding requires structure and evidence.
A useful engineering-intelligence system needs to know more than a file name or visual resemblance. It needs explicit entities, coordinate frames, features, measurements, relationships, and traceable support for what it claims.
02
Extract features
Identify measurable features such as holes, slots, copes, notches, axes, profiles, and other manufacturing evidence.
03
Map relationships
Represent orientation, adjacency, alignment, membership, connection evidence, and assembly context.
04
Retain evidence
Keep conclusions connected to the geometry and measurements that support them.
RESEARCH APPROACH
Geometry becomes structured physical knowledge.
GeometryLab's direction is deterministic first: establish what can be measured and represented explicitly before higher-level reasoning is asked to interpret it.
01
Ingest
Receive CAD-derived geometry and engineering context.
02
Normalize
Establish repeatable coordinate frames and comparable representations.
03
Recognize
Identify entities, shapes, features, and fabricated-member evidence.
04
Relate
Map geometric and assembly relationships between recognized entities.
05
Expose
Return structured results with measurements, warnings, confidence, and evidence.
WHY IT MATTERS
Reliable geometry can give both AI and people better engineering tools.
Engineers need to know which part details they can rely on and which still need checking. GeometryLab investigates the evidence and measurements that could help CADSync explain its findings, reduce repetitive review, and support clearer drawing and material-list preparation.
AI
Physical grounding
Give AI systems explicit geometry and relationships to reason from instead of relying only on language or appearance.
PEOPLE
In-context understanding
Support tools that explain what a feature or relationship means while the user is doing real engineering work.
AUTOMATION
Reliable process logic
Enable automation to act on structured engineering facts rather than fragile screen-level assumptions.
INTEGRATION
Portable evidence
Make geometric results usable across CAD tools, engineering services, reports, APIs, and connected workflows.
CURRENT DEVELOPMENT DIRECTION
Focused on engineered-object recognition and fabricated geometry.
Current research explores recognition of structural and fabricated members, manufacturing features, geometric relationships, and shared software foundations for Stratic tools. Coverage remains limited; results from a narrow test case do not establish support for every shape, material, or production workflow.
SHAPES
Engineered members
Structural and fabricated member recognition built from measurable geometry and repeatable evidence.
FEATURES
Manufacturing detail
Holes, slots, copes, notches, miters, weld-related geometry, and other useful fabrication signals.
PLATFORM
SDK & API foundation
Shared contracts and services that make geometric intelligence usable by multiple Stratic engineering tools.
STRATIC SYSTEMS
Building capabilities behind CADSync and other Stratic systems.
The purpose of this internal program is practical: turn research into capabilities that can help people review engineering information and carry it into their next task.