How the Planting Plan Works

What it does, step by step

1. Starts from the real measured site. Pulls your LaunchPad layers — Parcel, Lawn, Concrete/Asphalt, Building, and Beds/Natural — and draws them to true scale.
2. Plants only inside the measured beds. Lawn, driveway and the house are drawn for context but never planted — it can't put a shrub on the driveway.
3. Lays a spacing grid (~2.6 ft on center) across the beds, keeping only points that fall inside the bed shapes. On big estates it widens spacing automatically so it stays legible.
4. De-grids the dots with organic jitter so it reads like a natural planting, not a robotic grid.
5. Edges get groundcover, interiors get shrubs. Bed borders → low groundcover/grass (Lantana, Muhly). Interiors → structure shrubs (Hibiscus, Ixora, Firebush, Bougainvillea, Plumbago), clustered into color masses so it looks designed.
6. A few specimens anchor it. Up to three well-spaced points become Foxtail Palm feature specimens, kept far apart.
7. Every dot is one plant — countable. That's where the plant quantities in your budget come from. A soft "heat" blob under the dots shows the color mass; a legend lists every species.

What's genuinely strong

• Grounded in real measurements at true scale — not a generic clip-art garden.
• Produces countable quantities that flow straight into the budget.
• Real design logic: groundcover on edges, shrubs massed in the interior, specimens spaced as features.
• Never plants outside the measured beds.

The honest gap — which species (the part to improve)

The placement is smart, but species selection is a fixed full-sun scheme. It does not yet:

• Read the property's own sun/shade maps — so it can't yet put shade-lovers in the shade and sun-lovers in the heat. It assumes full sun everywhere.
• Pull from the full plant database (22 plants tagged by sun need, plus Trefle/Perenual) — it draws from a small hardcoded list.
• Take client inputs — color preference, exotic vs. classic, low-maintenance, budget.
• Apply nursery rules — odd-number groupings (3s, 5s), companion planting, spacing by each plant's mature size instead of a fixed grid.

Where it should go

Drive selection from sun map × plant database × client preferences: cooler north/east zones get shade species, hot south/west zones get full-sun species; vary species from the full library, priced per plant; space by mature canopy size, group in odd numbers, rotate bloom color by season.

How this strengthens the render

The AI render gets better the moment selection does: instead of a generic "lush full-sun yard," the render is fed the actual chosen species placed in their actual bed zones — so the picture the client sees matches the plan and the quantities they're paying for. Plan accuracy → render accuracy. The placement engine already gives us the "where"; richer selection gives the render the "what."