Get Cited by ChatGPT From a Squarespace Site

The current ChatGPT surface has three citation moments. ChatGPT Search shows a list of source cards at the top of an answer when the user's query is web-grounded. Conversational citations appear inline inside a regular answer when the model decides to back a specific claim. "More sources" footers sit underneath an answer when the user expands to see what else was consulted. OpenAI extended the utm_source=chatgpt.com tag to the More sources surface on 13 June 2025⁹, which finally made that traffic visible in GA4. The conversational inline citations remain untagged.

The pattern that gets cited across all three surfaces is consistent. Search Engine Land's 2026 GEO research⁴ finds that AI engines lift from "clear, direct answers" at the top of a section, expanded with context. ChatGPT specifically favours pages with named-source attribution, recent dates next to claims, and one or two specific statistics over hedged generalities. A page that reads like a short technical brief gets quoted; a page that reads like a sales pitch gets read past.

The Squarespace-specific complication is that the platform's defaults are right but the field guidance is wrong. Squarespace ships the AI exclusion box unchecked², which is the correct setting for any owner who wants to be cited. The trap is the 2023-2025 wave of "protect your content from AI" posts that told designers to flip the box on, and a meaningful share of currently-live Squarespace sites still carry that legacy setting.

OpenAI's bots page¹ states each crawler's job in one sentence. GPTBot: "GPTBot is used to make our generative AI foundation models more useful and safe." User-agent string starts with Mozilla/5.0 ... GPTBot/1.3. Respects robots.txt. OAI-SearchBot: "OAI-SearchBot is used to surface websites in search results in ChatGPT's search features." User-agent identifies as OAI-SearchBot/1.3. Respects robots.txt. ChatGPT-User: identifies as ChatGPT-User/1.0, and per OpenAI's own documentation, "because these actions are initiated by a user, robots.txt rules may not apply."

The robots.txt asymmetry is the part the field guidance keeps getting wrong. ChatGPT-User does not follow robots.txt because the request was made on behalf of a human asking a real question — a user typing "find me a wedding photographer in Asheville who shoots film" and ChatGPT going out to fetch the candidate pages. The behaviour is consistent with how a browser handles a typed URL: a robots.txt rule does not stop a person from visiting a page they explicitly asked for. The implication for Squarespace owners: even if the AI exclusion box is on, ChatGPT-User will still reach your site — but OAI-SearchBot will not, and OAI-SearchBot is the bot that decides whether your URL appears in the source-card list above an answer.

The Squarespace exclusion checkbox² covers GPTBot but not OAI-SearchBot or ChatGPT-User directly — the help center's 26-bot list names GPTBot among the AI training crawlers. Since the box is unchecked by default, the practical state for most owners is: all three OpenAI bots can crawl. The 90-second audit lives in the unblock-GPTBot leaf.

The reason to install in this order is mechanical. Crawler access without passages produces visits but no extractable quotes. Passages without source-naming produce extractable quotes that the model decides are not safe to surface. Source-naming without entity wiring produces citations that the model decides are not safe to attribute. Entity wiring without a measurement loop produces growth you cannot see. The five layers are sequential because the model's confidence in your page compounds from one layer to the next.

The verification step matters because the Squarespace UI state can lag the actual robots.txt file by a refresh. Look at the live file, not the checkbox. The full per-bot walkthrough with screenshots lives in the unblock-GPTBot leaf; the breakdown of which OpenAI bot does what lives in the ChatGPT-User vs GPTBot leaf. The AI Crawlers hub covers the broader 26-bot Squarespace list and where the OpenAI bots sit inside it.

Search Engine Land's 2026 GEO guide⁴ reframes this as "start each section with a clear, direct answer. Then expand with context." The 134-167 word band is the practical implementation: long enough to carry a full thought, short enough to fit a citation card. On a Squarespace 7.1 blog template, the format ships natively — drop a Markdown block (or a regular text block), bold the lead, and let the rest flow as body. The full block-by-block walkthrough with the heading hierarchy patch lives in the content-format leaf.

The mechanism is a hallucination guardrail. AI engines weight pages with explicit attribution more heavily because the citation gives the model a way to defend the claim if a user asks where it came from. A page citing "the Squarespace Help Center on AI exclusion (2026-Q1)" is materially easier to lift from than a page citing nothing. Cite the manifest right rail style this hub uses — primary source name, publication, date — and link out where you have it.

The full schema graph and Code Injection placement pattern is in Pillar 3 (Schema & Structured Data for Squarespace, when shipped). The single-page version is below — paste into Settings > Advanced > Code Injection > Header.

 <script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Person", "name": "Your Full Name", "url": "https://yoursite.com/founder/", "jobTitle": "Founder", "sameAs": [ "https://www.linkedin.com/in/your-handle", "https://github.com/your-handle" ] } </script> 

Most ChatGPT traffic still arrives without a referrer header — conversational inline links remain untagged, and mobile-app traffic strips referrers entirely⁹. Expect GA4 to undercount. Pair the channel grouping with the manual query log and Squarespace's own AI Visibility tool (Core/Business sites test every 14 days; higher plans every 7 days³). The dark-traffic leaf ships the regex, the GA4 setup screenshots, and the spreadsheet template.

OpenAI's announcement⁵ is unambiguous on the launch list: "Starting today" with Etsy sellers in the United States, with over a million Shopify merchants including Glossier, Vuori, Spanx, and SKIMS coming soon. Stripe's parallel newsroom post⁷ confirms the Agentic Commerce Protocol is an open standard co-developed with OpenAI, which leaves the door open to other platforms — but does not promise any timeline. Wix Commerce, Squarespace Commerce, BigCommerce, and WooCommerce are absent from both posts.

The product-feed specification⁶ describes the data shape OpenAI accepts: required fields including item_id, title, description, url, price with ISO 4217 currency, image_url, seller_name, seller_url, target_countries with ISO 3166-1 codes, plus checkout-eligibility flags. The submission mechanism for merchants outside the launch partners is not publicly documented as of mid-2026 — the path today is: be on Shopify or Etsy, or wait.

A Lawrence Hitches analysis⁸ of 100 brands across July 2024 to March 2026 tracked 340,000 ChatGPT sessions and $690,000 in attributed revenue, with year-over-year session growth of 19x — but flagged that GA4 referral numbers are "the floor, not the ceiling" because conversational inline links remain untagged, ad blockers strip referrers, and mobile-app traffic shows as direct. The aggregate conversion rate across the dataset was 0.9%, with category swings from 8.0% (printer supplies, high commercial intent) down to 0.01% (bridal, top-funnel research).

The Squarespace-side measurement layer is the AI Visibility tool inside the SEO panel³. Core and Business plans test prompts every 14 days; Plus, Advanced, and Commerce plans test every 7. The tool tracks branded prompts (does ChatGPT mention your brand?) and non-branded prompts (do you appear when someone asks about your category?). It is a useful signal but does not address the citation chain itself. The dark-traffic leaf ships the GA4 regex, the manual query log template, and the workarounds for measuring conversational-link traffic.