No black boxes. Pure math.

6-Orientation Checking

Every item is tested in all 6 possible orientations (3 axes × 2 rotations per pair). We never assume your couch has to stay upright — if tilting it on its side lets it through the door, we find that.

For each orientation, we compute per-axis clearances: width-across-opening, height-through-opening, and depth-along-passage. The orientation with the best minimum clearance wins.

Safety Buffers

Real-world moving isn't pixel-perfect. We deduct a safety buffer from every opening to account for door frames, wall fixtures, handles, and the reality of maneuvering large objects.

Default buffers: Doorways 38mm (1.5"), Hallways 25mm (1"), Stairs 51mm (2"), Turns 38mm (1.5"), Containers 2mm. Switch between Standard, Conservative (1.5×), or Exact (0×) modes.

Deterministic Math

The fit/no-fit verdict comes from pure geometry — bounding-box comparisons, sweep arc calculations for turns, and volumetric packing for containers. No language model touches the math.

AI parses your query (what item, what space, what dimensions). Then deterministic solvers compute the answer. Same inputs always produce the same result.

Confidence Levels

Every verdict comes with a confidence rating — HIGH, MEDIUM, or LOW — so you know exactly how much to trust the result.

HIGH: all dimensions user-provided and margin > 2". MEDIUM: database dimensions or moderate margins. LOW: estimated/default dimensions or margins under 0.5".

Graduated Verdicts

We don't just say "fits" or "doesn't fit." Five verdict tiers tell you exactly how tight things are, so you can decide whether to attempt it or find an alternative.

Tested Against 88+ Scenarios

Our solver suite is continuously tested against a ground-truth library of known fits and known fails — real furniture through real openings with verified outcomes.

Tests cover doorways, hallways, turns, containers, elevators, and rooms. Edge cases include exact-fit thresholds, orientation selection, and cross-constraint combinations.