Every fact, traced to its source.
When you open a page, PlantKey isn’t fetching data from anywhere — it’s already built in. We gather open data on a schedule, match it into one database, and build every plant page ahead of time, so each one loads in a few kilobytes and every fact links back to where it came from.
Source registry What we store vs. link · always check current license
Facts (bloom time, native status, height) aren't copyrightable, but licenses, API terms, and database rights still apply — and write-ups, photos & curated maps are copyrighted. We prefer public-domain / open sources, pull from official bulk endpoints, attribute per field, and link out to the rest. This isn't legal advice — verify each source before launch.
Data quality verified at build · every fact cited or omitted
454,903 facts cited across 31,673 taxa, from 32 open datasets. If any fact is missing its source, the build stops — so everything you see has been checked, and gaps are honest blanks, never guesses.
Every field, traced
Example: Asclepias syriaca — each value on a plant page carries its origin.
Full citations cite us, and cite our sources · every link verified
The per-field credit on every plant page links back to one of these. We cite the primary dataset (and link its DOI / deposit) rather than reproduce any copyrighted write-up — so you can trace, verify, and re-cite each fact at the source.
- USDA PLANTSUSDA, NRCS. The PLANTS Database. National Plant Data Team, Greensboro, NC. Structured DwCA deposit on Zenodo (public domain, US government). doi.org/10.5281/zenodo.17903503 →
- GBIFGBIF.org. Global Biodiversity Information Facility — occurrence download (per-record CC). Used to generate our own range maps. www.gbif.org →
- World Flora OnlineWFO (2024). World Flora Online. Published on the Internet (CC BY). www.worldfloraonline.org →
- US-RIISSimpson, A., et al. (2022). United States Register of Introduced and Invasive Species (US-RIIS), version 2.0. U.S. Geological Survey data release (public domain). www.sciencebase.gov/catalog/item/62d59ae5d34e87fffb2dda99 →
- USACE NWPLU.S. Army Corps of Engineers (2022). National Wetland Plant List (NWPL), 2022 edition (public domain). wetland-plants.sec.usace.army.mil/ →
- USA-NPNUSA National Phenology Network. Phenology Observation Portal: Magnitude Phenometrics (CC BY 4.0). doi.org/10.5066/F78S4N1V →
- Herbarium phenologyPark, D.S. et al. (2023). Herbarium-Derived Phenological Data in North America. Zenodo (CC BY 4.0). 2.3M digitized specimens scored for open flowers; we use flowering-month distributions per species as a bloom gap-filler. doi.org/10.5281/zenodo.8323155 →
- TRY (database)Kattge, J. et al. (2020). TRY plant trait database — enhanced coverage and open access. Global Change Biology 26:119–188 (CC BY 4.0). doi.org/10.1111/gcb.14904 →
- TRY (global spectrum)Díaz, S. et al. (2016). The global spectrum of plant form and function. Nature 529:167–171. (TRY File Archive ID 81, DOI 10.17871/TRY.81.) doi.org/10.1038/nature16489 →
- Mature width — open-grownMcPherson, E.G., van Doorn, N.S., Peper, P.J. (2016). Urban Tree Database. Forest Service Research Data Archive, RDS-2016-0005 (U.S. Government public domain). Open-grown crown width derived as the 95th-percentile of measured average crown diameter (AvgCdia) per species, ≥10 trees. doi.org/10.2737/RDS-2016-0005 →
- Mature width — wildJucker, T. et al. (2025). The global spectrum of tree crown architecture. Nature Communications. Data: Tallo, Zenodo DOI 10.5281/zenodo.14217401 (CC BY 4.0). Wild/forest-grown crown width derived as the 95th-percentile of measured crown_diameter_m per species, ≥10 trees. doi.org/10.5281/zenodo.14217401 →
- Frequently grows withSabatini, F.M. et al. (2021). sPlotOpen – An environmentally-balanced, open-access, global dataset of vegetation plots. Global Ecology and Biogeography 30:1740–1764. Data: iDiv Data Repository, DOI 10.25829/idiv.3474-40-3292 (CC BY 4.0). Per species, the catalog species most often co-recorded in the same U.S. vegetation plot (conditional frequency over the 19,098 U.S. plots; focal species ≥20 plots). A field co-occurrence signal, not a planting prescription. doi.org/10.25829/idiv.3474-40-3292 →
- NatureServeNatureServe (2024). NatureServe Explorer. NatureServe, Arlington, Virginia (CC BY — public ranks). explorer.natureserve.org/ →
- Federal ESA statusU.S. Fish & Wildlife Service. Environmental Conservation Online System (ECOS) — Threatened & Endangered Species (TESS). Public domain (U.S. Government work). ecos.fws.gov/ecp/ →
- Caterpillar host countWarren, R. J. II (2026). Phylogenetic and geographic distance from native plants shape herbivore associations with non-native plants. Dryad (CC0). LepTot = total Lepidoptera per genus from published host-use records (the Tallamy host-use compilation). doi.org/10.5061/dryad.1rn8pk17s →
- Named caterpillarsRobinson, G.S., Ackery, P.R., Kitching, I.J., Beccaloni, G.W., Hernández, L.M. HOSTS — a Database of the World's Lepidopteran Hostplants. Natural History Museum, London (CC0). Filtered to Nearctic records. doi.org/10.5519/havt50xw →
- GloBIPoelen, J.H., Simons, J.D., Mungall, C.J. (2014). Global Biotic Interactions: an open infrastructure to share and analyze species-interaction datasets. Ecological Informatics 24:148–159. (Caterpillar-count gap-fill where Warren lacks a genus.) doi.org/10.1016/j.ecoinf.2014.08.005 →
- Wildlife & visitorsGlobal Biotic Interactions (GloBI), Interpreted Data Products. Zenodo, DOI 10.5281/zenodo.5708970 (CC0). Per-species distinct counts of birds/mammals recorded eating the plant and adult Lepidoptera recorded visiting its flowers; vertebrate-class and Lepidoptera-order filtered. doi.org/10.5281/zenodo.5708970 →
- Recorded bee visitorsNoori, S. et al. (2026). A curated and integrated dataset for exploring global bee-plant interactions. Scientific Data. Zenodo (CC BY 4.0). Taxonomy-harmonized from GloBI via BeeBDC + WCVP. doi.org/10.5281/zenodo.18303036 →
- Specialist beesSmith, C. et al. (2024). Pollen-specialist bee species are accurately predicted from visitation, occurrence and phylogenetic data. Oecologia. Data: Big-Bee-Network/Bee-Specialization-Modeling, Zenodo (CC BY 4.0); host-plant associations from Fowler & Droege. doi.org/10.5281/zenodo.10420917 →
- Nitrogen fixationWerner, G.D.A., Cornwell, W.K., Sprent, J.I., Kattge, J., Kiers, E.T. (2014). A single evolutionary innovation drives the deep evolution of symbiotic N₂-fixation in angiosperms. Nature Communications 5:4087. Data: Dryad (CC0). doi.org/10.5061/dryad.05k14 →
- Rooting depthFan, Y., Miguez-Macho, G., Jobbágy, E.G., Jackson, R.B., Otero-Casal, C. (2017). Hydrologic regulation of plant rooting depth. PNAS 114(40):10572–10577. Data: Dryad (CC0). doi.org/10.5061/dryad.mgqnk99bg →
- Mycorrhizal typeMeng, Yiming; Davison, John; Clarke, John et al. (2024). Data from: Environmental modulation of plant mycorrhizal traits in the global flora [Dataset]. Dryad (CC0). FungalRoot v2.0. doi.org/10.5061/dryad.x0k6djhq2 →
- Stress toleranceNiinemets, Ü. & Valladares, F. (2006). Tolerance to shade, drought, and waterlogging of temperate Northern Hemisphere trees and shrubs. Ecological Monographs 76(4):521-547. Appendix A, figshare (CC0). doi.org/10.6084/m9.figshare.3565671 →
- Documented human usesU.S. Department of Agriculture, Agricultural Research Service (2023). Dr. Duke's Phytochemical and Ethnobotanical Databases [Dataset]. Ag Data Commons (CC0). Historical ethnobotanical record (food/medicinal class, count, and specific recorded uses from the ETHNOBOT table) — not medical or dietary advice. doi.org/10.15482/USDA.ADC/1239279 →
- Growth strategy (CSR)Pierce, S. et al. (2017). A global method for calculating plant CSR ecological strategies applied across biomes world-wide. Functional Ecology 31:444–457. Derived via the StrateFy calculation from leaf area / SLA / LDMC; ported verbatim from the open-source CSRcalculator (Gaskin et al. 2025). doi.org/10.1111/1365-2435.12722 →
- Chromosome number (2n)ChromoDB — Source Data: chromosome counts from the IAPT/IOPB chromosome data series (Zenodo, CC BY 4.0). The somatic 2n count + recorded cytotype variation, joined by accepted binomial. We render the raw count; ploidy LEVEL (diploid/tetraploid) is not asserted (it needs a per-genus base number we do not reliably hold). This is the license-clear Zenodo deposit, not the CCDB live portal. doi.org/10.5281/zenodo.18343310 →
- Nutrition (edible part)U.S. Department of Agriculture, Agricultural Research Service. FoodData Central — SR Legacy [Dataset]. Public domain (U.S. Government work). Baseline composition per 100 g, matched to native edibles by a curated, hand-verified crosswalk (FDC carries no scientific-name join). fdc.nal.usda.gov/ →
- Climate niche (heat / rainfall)NOAA National Centers for Environmental Information. nClimDiv county climate normals, 1991–2020 (public domain). Realized climate niche derived as the median annual rainfall and warm-end (p90) mean temperature across a species’ U.S. county range (GBIF). A realized, sampling-biased niche — not a physiological optimum. www.ncei.noaa.gov/pub/data/cirs/climdiv/ →
- Extra common namesWikidata contributors. Wikidata: the free knowledge base — taxon common names (property P1843). Wikimedia Foundation (CC0 1.0). www.wikidata.org →
- Flower colourMcKenzie, P., Berardi, A.E., Hopkins, R. (2025). Delayed flowering phenology of red-flowering plants in response to hummingbird migration. Flower colours assigned by GPT-4V from iNaturalist community-science photos (~87% expert agreement); repo pmckenz1/flower_color_phenology (MIT). Rendered as a labelled image-analysis field, confidence-gated. github.com/pmckenz1/flower_color_phenology →
- PhotosiNaturalist — CC-licensed observation photographs (CC0/BY/BY-SA/BY-NC/BY-NC-SA only), credited & licensed per image; Wikimedia Commons (Wikidata P18) gap-fills with one free-licensed image (CC0/BY/BY-SA/PD) for species iNat missed. Each photo keeps its OWN per-image licence — NOT relicensed CC-BY-SA. Some are NonCommercial (CC BY-NC / BY-NC-SA), which suits a free/non-commercial atlas; anyone reusing the images commercially must re-filter the NC photos out. www.inaturalist.org →
- DescriptionsWikipedia, The Free Encyclopedia (CC BY-SA 4.0). Each description links to the exact source article. www.wikipedia.org →
Citing PlantKey? PlantKey (2026). plantkey.org. Compiled open plant dataset, CC BY-SA 4.0. Please also cite the underlying source(s) for any fact you reuse.
How Ecological Value is scored
The 0–5 headline on each plant page is a keystone (host) value, built from open records of which animals depend on the plant — never guessed. It counts caterpillar food for butterflies & moths and pollen for specialist bees; it is not generalist nectar. The headline is the highest of its three parts, so a plant that’s outstanding for even one still scores high — an oak (caterpillars) and a sunflower (specialist bees) both reach 5, for different reasons. A low headline means “not a keystone host,” not “low value”: generalist nectar plants get a separate, labelled derived Flower-visitor value (below) so they’re not sold short.
Caterpillar species whose larvae eat the genus — food for birds and the wider web. Keystone count from Warren II 2026 (CC0; Tallamy host-use records), gap-filled by our GloBI derivation.
Bee species that feed their young only this plant’s pollen (pollen specialists), recorded on the genus. From Smith et al. 2024 (CC BY 4.0).
A native-milkweed override, since milkweed hosts few caterpillars and zero specialist bees — without it the keystone monarch plant would score ~2.
Authored Asclepias rule (our fact overlay).
A second, labelled axis shown beneath the keystone headline — the generalist nectar dimension the keystone score deliberately excludes. It bands the distinct native & managed bee species recorded visiting the plant (Noori et al. 2026, CC BY 4.0, via GloBI), lifted by any authored pollinator rating. It is recorded visitor breadth — observations, not a nectar-volume measurement — so it is banded coarsely and labelled “derived,” never shown as a direct fact. Bands track the observed distribution across the catalog (≈quartiles of visitor breadth):
Genus counts assume the native members carry the value local insects evolved with, so a non-native member is capped after the max:
"Invasive here" isn't baked — the live overlay adds a local "invasive in <state>" flag at request time.
Open interaction coverage is partial — absence is not zero. When no component is sourced the score is omitted entirely, never a rendered 0. Keystone scoring is genus-level (per-species nuance isn't captured). Caterpillar counts are the North-American keystone figures (Warren II / Tallamy), gap-filled by GloBI and gated by nativity. Generalist nectar isn't part of the keystone number — a nectar favourite like Echinacea lands low on keystone, but its separate derived Flower-visitor value (above) captures that. That axis is recorded visitor breadth, not nectar volume — well-studied plants accrue more records, so we band it coarsely and label it derived.
How derived cold-hardiness works
USDA gives an explicit hardiness zone for only ~6% of taxa. For the rest we derive a conservative floor — "hardy to at least zone N" — by crossing each plant's recorded U.S. range with the USDA Plant Hardiness Zone map. It's labeled derived on the page, never shown as a USDA fact, and USDA's own zone always wins where it exists.
For each state the plant grows in, we take that state's warmest zone — the plant is somewhere in that state, so it clearly survives at least that mild a winter. Across its states we then keep the coldest of those guarantees. The result is a true lower bound on its hardiness.
State data doesn't say where in the state. Taking a state's coldest corner would wrongly call a frost-tender coastal plant hardy in the mountains — a guess we can't back up. Warmest-per-state never overclaims; it can only under-state, which "at least" makes honest.
It's a floor, not a rating — a cold-hardy plant in a big state can read about 2 zones milder than its true limit. Plants confined to frost-free tropics (zone 12–13) are omitted rather than labeled meaninglessly. Finer county/point range data — on the roadmap — would tighten it. Source map: USDA PHZM (CC BY) × public-domain Census ZCTA centroids.
Update cadence
Open & cited
PlantKey's compiled factual dataset is published CC BY-SA 4.0; the code is MIT. Photos retain their original per-image licenses and credit, and every value on every page links back to the source it came from — nothing is gated or guessed.