How the site is found — SEO & AEO

Two kinds of machine look at this site now, and it is built for both.

• Search engines (Google and the rest) crawl the pages and list them, so a person searching for a country, a letter or a banknote can land on the page that has it.
• Answer engines — ChatGPT, Claude, Perplexity and their crawlers — read the site to answer with it. They are given a plain-text summary of the whole archive written for exactly that purpose.

Both are fed automatically by the build. There is no separate SEO tool, no plugin, and nothing to keep up to date by hand.

What the site publishes, and where

You can open every one of these yourself; they are ordinary public addresses.

• /llms.txt — the site in one page for an answer engine: what the archive is, its sections, and every country with its photograph and gallery counts.
• /llms-full.txt — the deep version: the full text of every journal letter, the banknote library, and the gallery inventory.
• /sitemap.xml — every public page, for search engines.
• /robots.txt — what crawlers may look at.

All four are rebuilt every time the site is published, from the same data the pages come from. They cannot fall behind the site, and no one has to remember them.

Each page also carries, generated from its own record: a description, the title and picture used when the page is shared, a canonical address, and — on the home page, gallery pages and journal entries — a small machine-readable block describing what the page is (a photo gallery and how many photographs, a journal entry and its date).

What you do

Routinely, nothing. The list below is the whole of it.

1. After a big change goes live — a batch of new albums, a run of new letters — open www.traveller.org/llms.txt in a browser and glance at it. The counts should have moved. That one page is the quickest proof the whole machine ran.
2. If you have added a brand-new page you specifically want in Google sooner rather than later, ask for it once in Search Console: URL Inspection → Request Indexing. That is rationed to roughly ten a day and is the only place asking helps.
3. Otherwise leave the reports alone. See below.

The one thing not to do

Do not press anything on the “Page with redirect” or “Duplicate” reports, or on the old-site addresses under “Crawled — currently not indexed”. Those reports emptying out is the goal — they are the old site being forgotten, exactly as intended — and asking Google to index those addresses works against it. The same goes for the whole second property in Search Console — the http:// one, the plain (non-https) www address of the old site: it is a museum, and it is meant to decay quietly.

The owner’s manual carries the full reading of those reports, and the reason each one looks alarming while being the machine working correctly.

Do not block the AI crawlers

The site deliberately blocks no answer-engine crawler. Publishing llms.txt and then blocking the crawlers that read it would cancel out the site’s entire answer-engine presence — it is the single change most likely to make the archive invisible to AI answers. If anyone ever suggests adding those blocks, that is the trade being made.

What is deliberately not findable

Some things are kept out on purpose, and their absence is not a fault:

• Hidden albums and friends-only photographs — absent from the sitemap, absent from llms.txt and llms-full.txt, absent from every count, and marked so search engines do not list them. They open only for someone holding the address or signed in as a friend.
• The admin pages and these guides — closed to crawlers in robots.txt.

If something looks wrong

• A new album or letter is missing from /llms.txt — it is almost always that the change has not been published yet. Publish, wait for the site to rebuild, and look again.
• Google is showing an old address — that is expected for a while. Old addresses forward to the right page, and Google works through them at its own pace; nothing needs pressing.
• Something is listed that should be hidden — that is worth reporting, because it means an album’s hidden setting is not what you think it is, not that the search machinery misbehaved.