What a guild is
A guild is a small group of plants arranged in layers around a central one, so that each layer does a job the others cannot: something covers the soil and holds moisture, something fixes nitrogen, something flowers when the others do not. The idea comes from forest gardening, where the model is a woodland edge rather than a flower bed.
This tool assembles one from facts we can cite, for any of the 2,169 plants in the catalogue.
How it builds the suggestion
Pick a target plant and optionally a ZIP code, and we assemble three layers:
- Ground cover — plants whose derived layer is ground cover, meaning under 1.5 ft and genuinely spreading across the soil surface rather than merely short.
- Bloom overlap — plants whose USDA bloom period overlaps your target's, so the planting has something in flower across a longer stretch of the season.
- Nitrogen fixers — plants USDA records as fixing nitrogen.
If you give a ZIP code, every suggestion is also filtered to your hardiness zone. Layers are derived from growth habit and mature height by a fixed rule, so the same plant always lands in the same layer.
What it refuses to suggest
Two kinds of plant are excluded from every layer, both flagged by USDA: those recorded as allelopathic — they release compounds that suppress neighbouring plants, which is the exact opposite of a support species — and those carrying an invasive or noxious status. That is 162 plants the builder will never recommend as a companion.
If the plant you chose as the target carries one of those flags, we tell you rather than quietly substituting something else. Your choice stays yours. We wrote up the allelopathy data in full — 41 of the 57 flagged plants are routinely sold, including sugar maple.
Bloom overlap is not a pollinator claim
This is worth being precise about. The third layer matches on USDA's Bloom Period, which is a statement about when a plant flowers. It is deliberately not a claim about pollinator value. No open dataset rates pollinator value reliably, so we do not offer it, and we removed the field from the database rather than estimate it. Overlapping bloom is a real and useful structural property; treating it as a pollinator score would be us inventing a fact.
What it cannot do
It does not publish pairwise companion-planting relationships — the "plant basil with tomatoes" tradition. We looked hard for a usable source. Companion planting is overwhelmingly a vegetable-garden practice, and USDA PLANTS does not characterise cultivated vegetables at all: crops like basil, carrot and lettuce return no trait data, while wild and native species return around eighty fields each. The open datasets we could license cover those same vegetables, so almost none of it would apply to the native flora in this catalogue. Rather than import advice we cannot ground, we left it out.
It also has no knowledge of your soil, your drainage or how much room you actually have. Treat the output as a structurally sound shortlist to check against your own site, not a planting plan.
Common questions
Why are there only a few suggestions per layer? Because we would rather show a short list that passes every filter than pad it out. If a layer comes back thin, it usually means your target's zone or bloom window is unusual.
Can I use a vegetable as the target? Not usefully — USDA does not characterise cultivated vegetables, so they are not in the catalogue.
Why does it suggest a plant taller than my target? Layers are assigned by absolute height, not relative to your choice. If your target is a ground cover, the structural layers above it will legitimately be taller.