Layer 1
Bronze
The seed.
New in this layer
RAW SOURCE DATARaw, last-known state per source. Updated on every scrape; never deleted on absence alone.
About/The reference
Plantaerium is a curated plant reference database from Realmdrop. We consolidate cultivar identity, attributes, and provenance into a clean, deduplicated record. We don't sell plants, hold inventory, or publish prices.
Scope
Why precision matters
“A canonical record isn’t a vote count. It’s a stand on what each cultivar actually is.”
Every field carries a source. Every disagreement carries a paper trail. The work isn’t storing more data — it’s deciding what counts.
Architecture
Every public-facing record passes through four layers. Each layer is conservative — supplier-side changes that conflict with your view never silently overwrite it.
Layer 1
The seed.
New in this layer
RAW SOURCE DATARaw, last-known state per source. Updated on every scrape; never deleted on absence alone.
Layer 2
The seedling.
New in this layer
SYSTEM MERGECanonical record per cultivar. Field-level provenance preserved — you can see exactly which source contributed what.
Layer 3
The leaves.
New in this layer
EDITORIAL LAYERSilver plus our editorial overrides for fields where supplier data conflicts or is incomplete. Every override is logged.
Layer 4
The flower.
New in this layer
YOUR OVERRIDESGold plus per-tenant customizations — your private corrections layered on top of ours, visible only to your account.
Each layer builds on the last. You always see the most refined record your tier allows.
The thread
Tracing one cultivar through every source.
Workflow
A small curation team works the queue every day — reviewing high-stakes matches, opening overrides for the conflicts that matter, and pruning images that drift off-brief.
A trigram resolver scores supplier listings against the canonical entity index. High-confidence matches auto-link; the rest queue for review.
Field-by-field merge with deterministic precedence (longest description wins; tightest hardiness range wins; etc.) plus full source attribution.
A curator reviews each entity surfaced by the queue or QA reports; conflicts get a Gold override with a written reason.
Why this matters
“Every record names a living thing — bred by someone, grown by someone, remembered by someone.”
The catalog is built to be worthy of that.
FAQ
Questions specifiers and integrators ask the most.
Each supplier is scraped on a per-source schedule (most weekly, some daily). Silver consolidation runs after every supplier finishes; Gold edits land instantly. The home page shows the latest update timestamp.
The silver consolidator picks deterministically per field — for example, the longest description wins, the tightest hardiness range wins. When the result is wrong, a curator opens a Gold override and the rule's reasoning is captured in our audit log.
Yes. "Tres Amigos® (Mincautri)" and the bare 'Mincautri' clone always merge into one canonical entity, with the trade name as the consumer-facing title.
Subscribers in our white tier can layer their own per-tenant overrides without affecting the public record. For cross-the-board corrections, drop us a note via the contact page — we treat well-supported corrections seriously.
Each image is content-hashed at ingest, not URL-tracked. We license imagery from supplier catalogs that allow it; curators can pin or unpin images per cultivar to keep the gallery on-message.
Yes. Send us the cultivar name (plus any supplier or breeder reference you have) via the contact page and we'll add it to the curation queue. We prioritise requests from active subscribers, but the queue is open to everyone.
Yes — we actively welcome new sources. Get in touch and we'll walk you through the supplier onboarding flow (an exported CSV, an API feed, or a scrapeable catalog all work). Your data flows into bronze, merges into the canonical record, and stays attributed back to you in the provenance trail.
Where to next?
Open the catalog, inspect the API surface, or talk to us about subscriber access.