Gallery/Live trajectory/Autodesk Revit 2026
Relocating a garage by mirroring
A residential plan reworked so the garage moves left to right: split the continuous walls, copy the portion clear rather than move it against its constraints, mirror it on the Y axis, return it on a matched reference point, then rebuild the walls, circulation, doors and room tags it disturbed.
Click this panel, then ↑↓ to step. Shift jumps 5, Home/End to the ends.
Every mark is a recorded action. Narration was captured on a separate stream and re-aligned at import, delayed by the -57ms between the first video frame and the first audio sample.
Click any block to jump the clip to the moment it was said.
23 blocks covering 344 actions. Transcribed with whisper.cpp large-v3-turbo, then corrected by hand against the action stream — the raw pass misheard several Revit commands as near-homophones.
Garage relocation — assessor rubric
Scored against the twelve-criterion rubric supplied with Project B Task 1, on that rubric's own 1-4 scale. The result is 44 of 48, inside the 43-48 band the rubric calls Excellent. Nine criteria took full marks; four did not, and they cluster around the same thing — dimensions and scale, overall footprint, minimal change, and structural framing. Relocating the garage by mirroring a split portion is correct and was reasoned through aloud, but it moved more of the plan than the brief asked for, and the closing sequence is a series of corrective nudges to pull the area back: 5'5" off the garage, 3 feet off the office and bathroom, 9 inches off the living room. Scores and level descriptors are the assessor's, transcribed unchanged. The evidence links are drawn from the capture itself, and each jumps to the action it cites.
Garage Relocation
4 / 4 8.33% weightIs the garage relocated to the required position, with no unintended changes elsewhere?
Garage is relocated to the specified/new location exactly as required, with no unintended changes.
Site/Plan Relationship
4 / 4 8.33% weightDoes the new garage location integrate with the site, footprint, setbacks and surrounding elements?
New garage location integrates correctly with the site, building footprint, setbacks, and surrounding elements.
Dimensions & Scale
3 / 4 8.33% weightAre the garage dimensions and architectural scale accurate against the reference diagram?
Dimensions and scale are mostly accurate.
Orientation & Access
4 / 4 8.33% weightAre orientation, driveway access and vehicle circulation logical and resolved?
Garage orientation, driveway/access, and vehicle circulation are logical and clearly resolved.
Building Coordination
4 / 4 8.33% weightIs the garage coordinated with adjacent walls, doors, circulation and other Revit elements?
Garage is properly coordinated with adjacent walls, doors, circulation, and other Revit elements.
Overall Footprint / Area
3 / 4 8.33% weightHow closely does the resulting building area match the original plan?
Minor variation, less than 5%.
Compliance to Constraints
4 / 4 8.33% weightWere the stated environmental and site constraints respected?
100% complied.
Minimal Change Objective
3 / 4 8.33% weightWas the result reached with the smallest set of changes?
Minor changes.
Revit Model Quality
4 / 4 8.33% weightAre elements correctly placed, aligned and organised, with no stray geometry left behind?
Model elements are correctly placed, aligned, and organized; no unnecessary geometry is introduced.
Reference Diagram Compliance
4 / 4 8.33% weightDoes the final plan follow the requirements and bye-laws in the provided diagram?
Final Revit plan clearly follows all relevant requirements/bye laws from the provided diagram while implementing the garage relocation.
Presentation & Clarity
4 / 4 8.33% weightIs the final plan legible, with the relocated garage easy to identify and evaluate?
Final plan is clear, legible, and makes the relocated garage easy to identify and evaluate.
Structural Framing / Configuration
3 / 4 8.33% weightWhat happened to the structural framing as a result of the relocation?
Minimal changes but structurally good.
Excellent by the rubric's own banding. The approach is sound and the constraint reasoning is explicit — the sunroom is held to the sun path, and the portion is copied rather than moved so the upstream constraints are never fought. What holds it short of full marks is scope: the footprint moved, and had to be walked back.
Pairs come from the action space: every control that was available and not chosen forms one (preferred, rejected) pair against the control the expert took.