Methodology

How this plant database is built — what data we use, where it comes from, what's human-curated vs. AI-assisted, and what to trust it for.

What this is

A free, openly-accessible plant database covering 1,972 varieties across 30 categories, with zone-specific planting calendars, companion-planting data, nutrition data, and growing guides. Maintained by Brian L of Wind River Greens, a small microgreens farm in Milton, Georgia.

The database combines structured data sourced from authoritative third parties (extension services, USDA, breeder catalogs) with AI-assisted editorial content (growing guides, troubleshooting, companion guides) that's anchored in those real source materials. The "AI disclosure" section below explains exactly which content is which.

Data sources

Variety identity (name, scientific name), growing characteristics (zones, sun, soil, height), and other structured fields come from four authoritative sources. Where a variety appears in multiple sources, the data is reconciled in favor of the more cultivar-specific reference. Source attribution is shown on each variety page under the "Sources & References" section.

SourceUsed forVarieties cited
NC State Extension

Plants Database (NCSU)

Growing characteristics (light, soil, height, hardiness zones), pests, diseases481
Johnny's Selected Seeds

Breeder product data

Cultivar-level days-to-maturity, disease resistance, plant size, growth habit799
USDA FoodData Central

USDA Agricultural Research Service

Nutrition data per 100g (calories, macros, vitamins, minerals)1,036
Missouri Botanical Garden

Plant Finder

Supplementary deep content for select ornamentals11

Coverage: 1,822 of 1,972 varieties (92%) have at least one external citation. The remainder are typically niche ornamentals or houseplants not yet present in these public datasets.

How planting calendars are calculated

Planting calendars are formula-based, not AI-generated. For each variety + USDA hardiness zone combination, the calendar is computed from:

  • NOAA frost-date data per USDA zone (average last spring frost, average first fall frost)
  • Variety days-to-maturity from the breeder source (typically Johnny's)
  • Crop classification (warm-season / cool-season / perennial / indoor-only) determining sow-window logic

Indoor-only categories (houseplant, succulent) are correctly excluded from the calendar — they have no seasonal planting window. The current dataset includes 0 varieties with calendar entries spanning USDA zones 1–13.

AI-assisted editorial content

Some sections on each variety page are generated by AI (Claude Sonnet 4.6) with a structured editorial process designed to ground content in real sources and screen out AI-style filler. Disclosed here in full:

AI-generated sections

  • Troubleshooting (symptom → causes → fixes) — prompted with the variety's real pest/disease data from NC State Extension
  • Companion guide prose — prompted with the variety's real companion-plant records from our research database
  • Succession guide — prompted with the variety's actual zone range and days-to-maturity
  • Growing guide narrative (long-form) — prompted with NC State Extension content where available, otherwise generated from structured fields

Editorial process

  1. Draft generated from the variety's real source data (pests, diseases, zones, days-to-maturity)
  2. Self-critique pass: the model reviews its own draft for banned phrases, vague abstractions, and AI-style cadence, then rewrites
  3. Automated audit against a 47-phrase banned list ("tapestry," "delve," "navigate," etc.)
  4. Audit-flagged outliers regenerated on a stronger model (Claude Opus 4.7)

Not AI-generated

  • Variety identity (name, scientific name, taxonomy)
  • USDA hardiness zone ranges
  • Planting calendar dates (NOAA frost + days-to-maturity formula)
  • Nutrition data (USDA FoodData Central)
  • Companion plant identities (sourced from research; the prose around them is AI)
  • Source URLs (curated from real attribution)

Limitations

  • Days-to-harvest is an estimate. Real-world outcomes vary with local microclimate, weather year-to-year, soil, and seed source.
  • Zone ranges are guidance, not guarantees. Zone 7b in coastal Georgia and zone 7b in interior New Mexico are different growing conditions in practice.
  • This is a reference, not professional advice. For plant pathology, agricultural decisions, or commercial growing, consult your state extension service.
  • AI-generated text can be wrong. Despite source-grounding and audit, hallucinations are possible. If you notice an error, please report it.
  • Variety data was last refreshed in late April 2026. Future updates will be noted in the source-file timestamps.

Updates & corrections

Source data is refreshed periodically as the underlying datasets (NC State Extension, Johnny's, USDA FoodData Central) update. Each variety page shows a dateModified in its page metadata reflecting the most recent edit to that record.

Spot an error? Email info@windrivergreens.com with the variety name and the issue. Corrections are typically applied within a week.