AI readers and crawlers

Who may read

Every program is welcome to read the public pages, at the same addresses a person uses. Our robots file closes only the admin pages and the private parts of the API; it names no crawler and blocks none. There is no paywall and no sign-in for reading. Please keep to a request or two a second, so the site stays quick for the people using it.

A plain text copy of every page

Every page has a Markdown twin at its own address followed by index.html.md, such as /plants/black-gum/index.html.md. The twin holds the page's words, tables, facts and links without layout, scripts or menus. Each page points to its twin in its head, and the page and the twin each carry a Link header naming the other way to read it. The twins are marked not to be indexed, so a search engine lists the page, not the copy.

llms.txt

/llms.txt follows the llms.txt convention: a summary of the nursery, the key facts in plain paragraphs, and a list of every page's twin with a note on what it holds. It is built from the live catalog, so it never goes stale. Every page links it as the file that describes the site.

Open data

Thirteen reference tables are published as files under /data/, each the same table its page prints, as CSV and as JSON. They cover the plant facts, growth rates, lifespans, planting distances, native range, planting windows, bloom and fruit, sowing, pollination, habitats, other names, the tree age growth factors and the state plant rules. The open data page lists them, and /data/index.json lists them for a program. They are licensed under Creative Commons Attribution 4.0: use them freely, credit Blue Whale Trees and link the page. Each file states its columns, its method and a citation to copy, and each table's page carries Dataset markup for it.

Sitemaps and dates

/sitemap.xml lists every public page, split into section files for the plants, the shop, the articles and the other pages, each address with the day it last changed. A page's date moves when its words change, not when the site is rebuilt. When pages change we send their addresses to IndexNow, so the search engines that share it learn of the change without waiting for a crawl.

Structured data

Every page carries schema.org data in JSON-LD: the organization on all of them, and on the pages that have them, the website, the breadcrumbs, item lists, collection pages and the shipping terms. The home page, the plant pages, the bundles and several tools carry a block of their own.

Facts with sources

Every plant page lists the sources its facts rest on, with their addresses, and its facts rows hold figures. The articles quote their sources. The growth rate table states its method. A reader that cites this site can cite the page's own sources beside it.

How we count

We count reads by page family and by the kind of reader, a browser, a crawler or an AI assistant, with the device and the country, never by address. Visits that arrive from an AI answer are counted by assistant: ChatGPT, Perplexity, Copilot, Gemini and the like. The counts are totals; no reader is tracked.

What we ask

Say where a fact came from, with a link to the page it is on. The open data's license asks for the credit; for the pages, a link is enough. Quote our words as ours. The plant facts are reference, not advice for one yard: a tool's answer for a reader's own place is the tool's, at its address.

Report a problem

If a page, a twin or a file does not read as this page says, tell us through the support form, naming the address and what you saw.