Drawing Intake (AI)

PDF drawing-set intake workspace. Indexes and type-classifies sheets, extracts title-block project information, supports scale calibration, detects wall/bearing-wall/opening/lintel/beam/post/joist-direction/note/grid candidates from the PDF text layer and vector line work with confidence, evidence, and stated assumptions, and converts engineer-accepted items into a draft layout for the Structural Schematic Designer with full source traceability. Engineer markups can be dragged, resized, and reclassified; full raster computer vision/OCR remains scaffolded, not implemented.

Every detected item is an assertion that requires engineer review - nothing is applied to project data or the schematic model without explicit acceptance. Detection on arbitrary uploaded PDFs is limited to text-layer signals and vector line-pair heuristics (no raster OCR/CV); estimated dimensions depend on scale calibration.

Use to triage incoming drawing sets and seed a draft schematic layout. Accept, edit, or reject every suggestion; verify scale before relying on estimated dimensions.