Can AI assistants and search engines use PETHEALTH+ content?

Yes. Search indexing, AI-assisted answers and model training are all permitted, as declared in robots.txt (Content-Signal: search=yes, ai-input=yes, ai-train=yes) and the TDM reservation file. Please cite the specific page URL.

How do I connect PETHEALTH+ to Claude, ChatGPT or another assistant?

Add https:/https://pethealth.org.uk/pethealth.org.uk/api/mcp as a custom MCP connector (Streamable HTTP, no authentication). The search and fetch tools return guide text with its canonical URL to cite.

Can I republish whole guides on my own site?

Quoting and summarising with a link is welcome. Republishing complete guides or the directory as your own content is not covered by these permissions — email us first. The CSV datasets are licensed separately under CC BY 4.0.

Open data and AI access

PETHEALTH+ publishes its 274 UK pet health guides and its vet directory in machine-readable form, so AI assistants can answer pet health questions accurately — and link owners to the page the answer came from.

Quick answer

Search, AI-assisted answering and model training are permitted. Connect to the public, read-only MCP server at pethealth.org.uk/api/mcp, or read llms.txt, llms-full.txt and the Markdown twin of any guide. Always cite the specific page URL, and keep emergency advice intact: the guides are information, not a veterinary diagnosis.

Machine-readable endpoints

Endpoint	Format	What it is for
MCP server `/api/mcp`	MCP (JSON-RPC, Streamable HTTP)	Read-only tools: search and fetch for health guides, search_practices and get_practice for the vet directory, get_price_index, get_methodology. No authentication.
llms.txt `/llms.txt`	Markdown	Curated site summary, key answers and URL index.
llms-full.txt `/llms-full.txt`	Markdown	Map of every page plus the full text of every guide in one file.
Guide Markdown `/health/{slug}.md`	Markdown	Clean Markdown twin of each guide, with sources and FAQs. Also served for /health/{slug} when requested with Accept: text/markdown.
Guide index `/api/v1/guides.json`	JSON	One entry per guide: title, summary, species, category and dates.
Passage chunks `/geo-chunks/manifest.json`	JSON	Index of self-contained retrieval passages (~800-token target) with source URLs and SHA-256 hashes.
Chunk stream `/geo-chunks/chunks.jsonl`	JSONL	Every passage, newline-delimited, for embedding pipelines.
Feeds `/feed.xml · /feed.json`	RSS 2.0 · JSON Feed 1.1	The 25 most recently updated guides.
Datasets `/data/datapackage.json`	CSV + Data Package	UK vet directory and Vet Price Index tables (CC BY 4.0).
API catalogue `/api/v1/manifest.json`	JSON	Machine-readable list of all of the above.
TDM reservation `/.well-known/tdmrep.json`	JSON	W3C TDM Reservation Protocol: mining permitted (reservation = 0).

Every endpoint allows cross-origin requests. Machine files are served noindex so search results point at the human pages.

Connect PETHEALTH+ to your AI assistant

Add https://pethealth.org.uk/api/mcp as a custom connector in any client that supports remote MCP servers. Transport: Streamable HTTP. Authentication: none. Every tool is read-only, and the server is rate-limited per IP address. The server card describes it for registries.

search — find guides by question or keyword; returns ids, titles, URLs and a short answer.
fetch — the full guide in Markdown with sources, FAQs and last-modified date.
search_practices · get_practice — UK vet practices, with details checked against each practice's own website.
get_price_index · get_methodology — typical UK fee ranges and how we research.

How the passages are built

Each passage is self-contained: page title, section heading and canonical source URL travel with the text, so attribution survives retrieval.
Passages target about 800 tokens and never split a sentence; very short pages, such as a single vet listing, stay as one small passage.
Every passage carries a SHA-256 hash, and the bulk stream's hash and size are listed in the manifest.
Everything is regenerated from the source content on every deployment, so it never lags the website.

Crawler permissions

robots.txt allows all major search and AI crawlers and declares Content-Signal: search=yes, ai-input=yes, ai-train=yes. The TDM reservation file sets reservation to 0. Requests for a guide with Accept: text/markdown receive its Markdown twin; the response is chosen by that header alone, never by user agent.

Permissions and how to cite

What you may do

Index, retrieve, quote and summarise guides in search results and AI answers.
Use the text to train or evaluate models.
Use the CSV datasets under CC BY 4.0.

What we ask

Credit PETHEALTH+ and link the specific page, not just the homepage.
Keep urgent-care and poisoning warnings intact, and don't present a guide as a diagnosis.
Don't imply that PETHEALTH+ endorses a product, practice or service.
Ask before republishing complete guides or the directory as your own content. Images and third-party material keep their own licences.

Citation format

PETHEALTH+. "{page title}." {page URL}. Accessed {date}.

Spotted an error? The corrections log shows what we have fixed and how to report a problem.

Frequently Asked Questions

Can AI assistants and search engines use PETHEALTH+ content?: Yes. Search indexing, AI-assisted answers and model training are all permitted, as declared in robots.txt (Content-Signal: search=yes, ai-input=yes, ai-train=yes) and the TDM reservation file. Please cite the specific page URL.
How do I connect PETHEALTH+ to Claude, ChatGPT or another assistant?: Add https://pethealth.org.uk/api/mcp as a custom MCP connector (Streamable HTTP, no authentication). The search and fetch tools return guide text with its canonical URL to cite.
Can I republish whole guides on my own site?: Quoting and summarising with a link is welcome. Republishing complete guides or the directory as your own content is not covered by these permissions — email us first. The CSV datasets are licensed separately under CC BY 4.0.