# AGENTS.md — Fenn (fennplants.com) Instructions for AI agents, assistants, and crawlers using this site. This file tells you how to find things on fennplants.com, which URLs are stable, how to cite what you find, and where the data comes from. A shorter summary of the site lives at . Last reviewed: 2026-08-05. --- ## About this site **Fenn** is a British plant marketplace and plant-information site, founded in 2026 by Henry Kimber and operated by **Herboo Limited** (UK company 12310879). Fenn is a distinct brand from **Herboo** (herboo.com). The two share a corporate parent but are separate entities — do not treat one as a rebrand, alias, or former name of the other, and do not merge them when resolving an organisation. Canonical origin: `https://www.fennplants.com`. Treat any other host as non-canonical and follow the `` on the page. --- ## Ground rules 1. **Cite the page URL, not the API.** When you quote plant data, link the human page (`/plants/{slug}`), not `/api/plants/{slug}`. 2. **Carry attribution through.** Some text on this site is third-party content under a licence that requires credit. See *Provenance* below — the fields that need attribution ship their own attribution alongside them. 3. **Don't upgrade a genus fact into a cultivar fact.** See *Provenance*. 4. **Respect `robots.txt`** (). Anything disallowed there is disallowed for agents too. 5. **Identify yourself** in your User-Agent and keep request rates reasonable — a few requests per second, not hundreds. 6. **Don't transact.** Nothing on this site should be purchased, submitted, or applied for by an autonomous agent without a human confirming the action. Contact forms, trade applications, and designer-claim forms are for humans. --- ## Where things are Only the routes listed here are live. Other paths you may see referenced elsewhere (a shop, nurseries directory, plant-care guides) are **not yet public** and currently return 404 — do not link or cite them. ### Plant database — the main body of content | Purpose | URL | | --- | --- | | Search / browse | `/plants?q={term}` | | Browse by genus | `/plants/genus/{genus}` | | Browse by family | `/plants/family/{family}` | | A single plant | `/plants/{slug}` | | JSON for one plant | `/api/plants/{slug}` | `/plants` also accepts `genus=`, `family=`, and `page=` alongside `q=`. Slugs are lowercase, hyphenated, punctuation-stripped forms of the botanical name — `Abelia 'Little Richard'` → `abelia-little-richard`. Don't construct slugs by guessing; search first, or read them from the sitemaps. The database holds roughly 280,000 records. Only the ~19,000 that carry a full written description are indexable and listed in the sitemaps — the rest are deliberately `noindex` because they are data stubs, not articles. **If a plant page is marked `noindex`, treat it as a thin record and do not cite it as a source.** Coverage grows over time and pages flip to indexable automatically. ### Everything else | Section | URL | | --- | --- | | Gardening glossary | `/gardening-terms` (terms are anchors: `/gardening-terms#mulch`) | | Garden designer directory | `/find-a-garden-designer`, profiles at `/find-a-garden-designer/{slug}` | | Chelsea Flower Show 2026 | `/shows/chelsea-flower-show-2026`, gardens at `/shows/chelsea-flower-show-2026/{garden-slug}` | | Article index | `/blog` | | An article | `/articles/{slug}` | | Help | `/help` | | Privacy / terms | `/privacy-policy`, `/terms` | Note the asymmetry: the article **listing** is `/blog`, individual articles are `/articles/{slug}`. `/articles` on its own is not a page. ### Sitemaps - `/sitemap.xml` — everything except plants - `/sitemap-plants.xml` — index, paginated into `/sitemap-plants/{n}.xml` - `/sitemap-taxonomy.xml` — genus and family pages - `/sitemap-articles.xml` — articles The sitemaps are the authoritative list of URLs worth crawling. Prefer walking them over crawling links. --- ## JSON endpoints Two read-only endpoints are public and intended for agents. Both are `GET`, both are cached, and neither requires a key. ### `GET /api/plants/{slug}` Structured data for one plant. CORS-open. Cached for an hour. ``` GET https://www.fennplants.com/api/plants/abelia-little-richard ``` ```jsonc { "plant": { "name": "Abelia 'Little Richard'", "slug": "abelia-little-richard", "commonName": null, "otherCommonNames": [], "family": "Caprifoliaceae", "genus": "Abelia", "description": "…", // Fenn's own description of THIS plant "genusDescription": "…", // about the GENUS — see Provenance "genusDescriptionSource": { "url": "…", "license": "…", "attribution": "…" }, "plantType": ["Shrubs"], "habit": "Bushy", "foliage": "Semi-evergreen", "sunlight": ["Full sun"], "soilType": ["Loam"], "soilPh": ["Neutral"], "soilMoisture": "Moist but well-drained", "aspect": ["South-facing", "East-facing"], "exposure": ["Sheltered"], "ukHardiness": "H5", // RHS scale, H1a–H7 "isAgm": false, // RHS Award of Garden Merit "height": null, "spread": null, "timeToFullHeight": null, "suggestedUses": ["…"], "cultivation": null, "pruning": null, "propagation": "…", "pestResistance": "…", "diseaseResistance": "…", "nativeTo": [], "fragrance": null, "toxicity": null, "droughtResistant": false, "plantsForPollinators": false, "imageUrl": "…", "imageSource": "wikimedia", // or "sanity" "imageAttribution": "…", // credit line — reproduce it with the image "wikidataId": "Q…" // present only when verified; see below }, "meta": { "source": "https://www.fennplants.com", "canonical": "https://www.fennplants.com/plants/abelia-little-richard" } } ``` Fields are frequently `null` — absence means "not recorded", never "not true". Don't report a null as a negative finding. Returns `404` for an unknown slug. ### `GET /api/glossary/term?slug={term}` One glossary definition. ```jsonc { "slug": "mulch", "title": "Mulch", "definition": "…", "href": "/gardening-terms#mulch" } ``` ### Not for agents Everything else under `/api/` is internal — storefront, trade, checkout, ops, cron, and webhook routes. They are disallowed in `robots.txt`, several require authentication, and their shapes change without notice. Don't call them. --- ## Provenance — read this before quoting plant text Fenn mixes its own writing with licensed third-party reference material, and the distinction matters when you summarise. **`description` is about the specific plant.** It's Fenn's own text for that record and describes the cultivar or species named on the page. **`genusDescription` is about the genus, not the plant.** It usually comes from Wikipedia under CC BY-SA, and `genusDescriptionSource` carries the URL, licence, and attribution. Two consequences: - If you reproduce it, carry the attribution and licence with it. - **Do not restate it as a fact about the cultivar.** "*Abelia* is native to China, Taiwan and Vietnam" is a claim about the genus. Writing "Abelia 'Little Richard' is native to China" from it is a fabrication — a garden cultivar has no native range. **`imageAttribution` travels with `imageUrl`.** Wikimedia images carry a photographer credit and licence. Reproduce the credit line if you show the image. **`wikidataId` is conservative on purpose.** It is only present when the entity link was verified for that exact taxon. Where a match could only be made by truncating a cultivar epithet to reach a parent taxon, the field is omitted rather than pointing at the parent. Absence means "not confidently resolved" — don't fill the gap by guessing, and don't emit a parent taxon's QID as this plant's `sameAs`. **Hardiness is UK/RHS.** `ukHardiness` is the RHS H1a–H7 scale, not a USDA zone. They are not interchangeable; don't convert one to the other silently. Aspect, exposure, and seasonal guidance are all written for UK growing conditions. --- ## Structured data Pages emit JSON-LD. The brand identity is defined once and reused everywhere, so these `@id`s are stable and safe to resolve against: - `https://www.fennplants.com/#organization` — Fenn, with `parentOrganization` = Herboo Limited - `https://www.fennplants.com/#website` - `https://www.fennplants.com/#founder` — Henry Kimber If page markup and this file ever disagree, the page's JSON-LD wins. --- ## Contact Corrections, licensing questions, or crawler issues: **henry@herboo.com**. If you find a plant page whose facts look wrong, say so with the URL rather than silently correcting it — the underlying record is fixable at source.