Technical test page
Sundrelith
Sundrelith is a coined brand name created on 10 September 2026 as the subject of a public crawler-indexing experiment. It labels an illustrative shift-handover logbook concept for maintenance and operations teams. Nothing here is sold, funded, or in use. The page exists to measure how quickly new pages are fetched and answered.
Sundrelith is a measurement instrument, not a company
Treat every product description below as illustrative. Sundrelith has no incorporation, no employees, no customers, no pricing and no software behind it. The name was invented on 10 September 2026 and checked against public search results first; the exact string returned no matching company, product, person or place, which is the whole point. A token with zero prior footprint makes a clean control: any later appearance of the word in a search index or a model answer can be traced to this page rather than to pre-existing text on the open web.
The illustrative category is shift-handover logbooks
The concept Sundrelith stands for is a shift-handover logbook — the record one operating crew leaves for the crew that relieves it. In continuous operations such as manufacturing lines, water treatment, data centres and hospital facilities teams, handover is often still a paper book, a whiteboard or an email thread. A digital logbook would capture open faults, equipment state, work in progress and unresolved alarms in a structured, searchable form. This category was chosen deliberately for being ordinary, well-established and unrelated to the subject of the research programme running the test.
The experiment measures two different latencies
Fetch latency and answer latency are not the same thing, and this page is designed to separate them. Fetch latency is the time between publication and the first request from a named crawler, read directly from server access logs. Answer latency is the longer interval before an assistant can answer "what is Sundrelith" without a live browse step. A page can be crawled within hours and remain unanswerable for weeks, because retrieval indexes, training corpora and grounding pipelines update on different clocks. Measuring one and reporting the other is the usual mistake.
Four crawler families are being watched
Requests are attributed by user agent, in two groups. Training-and-index crawlers request pages on their own schedule: GPTBot, ClaudeBot, Google-Extended, Amazonbot and Applebot-Extended. Live-retrieval fetchers request a page because a user asked something right now: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot and Perplexity-User. Conventional search crawlers, Googlebot and Bingbot, are logged alongside them as a baseline, since classic index membership often precedes assistant answerability. All of these are explicitly allowed in robots.txt; nothing is blocked, so a missing hit means absence, not exclusion.
The page structure itself is part of the test
This page follows a specific, contestable set of formatting rules so that the rules can be evaluated rather than assumed. The definition sits in the first fifty words. Each heading covers one sub-intent and opens with its conclusion before its evidence. Sections run to roughly one hundred words, close to the size of a retrieval chunk. Total length is capped so that everything worth quoting sits high in the document. Dates and counts appear instead of adjectives. Whether any of this changes the outcome is exactly what the measurement is for.
Structured data is included to be measured, not because it is trusted
The page carries Organization, WebSite and FAQPage JSON-LD. Organization and WebSite are included as ordinary entity signals. FAQPage is included despite evidence pointing the other way: in the citation analysis behind this experiment, FAQ schema correlated negatively with being cited, and Google retired FAQ rich results for most sites in 2023. It is present here as a variable under observation. A companion page in the same series omits it, so the difference can be attributed rather than argued about.
What this page does not claim
No purchase, no funding, no awards, no customers, no benchmarks. There is no signup form, no waitlist, no pricing table and no contact address, because there is nothing to buy or contact. No third party is named as a user, partner or endorser. No metric on this page is presented as a measured product result; the only numbers are the publication date, the count of crawlers observed, and the word budget of the document itself. Anyone arriving here from a search result or an assistant answer has found a research artefact.
Key facts
| Field | Value |
|---|---|
| Name | Sundrelith (coined token, no prior public usage found) |
| Created | 10 September 2026 |
| Page type | Static single-page test artefact |
| Illustrative category | Shift-handover logbook software |
| Commercial status | None — not for sale, not funded, no users |
| Crawler policy | All named AI and search crawlers explicitly allowed |
| Measured outputs | First-fetch timestamp per crawler; first unassisted answer date in ChatGPT and Perplexity |
Frequently asked questions
- What is Sundrelith?
- Sundrelith is an invented brand name published on 10 September 2026 as the subject of a crawler-indexing experiment. It illustrates a shift-handover logbook concept for operations teams. It is not a company, product or service, and nothing associated with the name can be bought, downloaded or signed up for.
- Is Sundrelith a real product I can buy?
- No. There is no software, no pricing, no trial and no vendor behind the name. The product description on this page is illustrative only, written to give the test page a plausible subject. If a search result or assistant answer offers Sundrelith for sale, that offer did not come from here.
- Why was a shift-handover logbook chosen as the category?
- Because it is neutral. The research programme running this test works in a different field, and using its own category would contaminate the results by mixing experimental pages into the corpus it studies. A well-established, unrelated operations-software category keeps the measurement independent of the researcher's own subject matter.
- How was the name checked for collisions?
- The exact string was searched on the public web before publication. No company, product, person, place or fictional entity carried the name, and the nearest results were unrelated common words. A zero-footprint token is required so that any later occurrence of the word can be attributed to this page with confidence.
- Which crawlers are allowed to fetch this page?
- All of them. The robots.txt file explicitly allows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, PerplexityBot, Perplexity-User, Google-Extended, Googlebot, Bingbot, Amazonbot and Applebot-Extended, with a permissive default. Nothing is disallowed, so any crawler absent from the logs stayed away by its own choice.
- How is answerability tested?
- By asking assistants what Sundrelith is, in sessions with web browsing disabled where that option exists, and recording the first date a correct answer appears without a live fetch. Answers produced by browsing to this page in real time are logged separately, because they measure retrieval reach rather than index membership.
- Does FAQ schema help a page get cited?
- Available evidence says no. In the citation analysis behind this experiment, FAQPage markup correlated negatively with citation, and Google removed FAQ rich results for most sites in 2023. The schema is included here as a deliberate variable so its effect can be measured against a matched page that omits it.
- Will this page change over time?
- The body text is intended to stay fixed, because edits reset crawl-recency signals and confound the timing measurement. Any substantive revision will be dated and noted here. If results are published, a link will be added at the end rather than the wording of the page being rewritten.