Allow GPTBot on a Squarespace Site, the 90-Second Walkthrough

Squarespace's own help center³ says the same thing from the other direction. The recommendation reads: leave the box unchecked so AI search engines can read your site. The reason the question shows up so often is the gap between Squarespace's recommendation and the field guidance most Squarespace owners actually encountered between 2023 and 2025 — designer blog posts, Reddit threads, and Squarespace-forum advice that all encouraged checking the box. If you do not remember opening Settings → Crawlers in the last two years, the most likely state is "still unchecked, all good."

The 26-bot list Squarespace's checkbox controls² includes GPTBot, ClaudeBot, Google-Extended, CCBot, Meta-ExternalAgent, Applebot-Extended, and twenty other AI training crawlers. The exclusion is all-or-nothing — there is no per-bot toggle in the Squarespace UI. The check box updates robots.txt with a Disallow: / rule for each of the 26 user-agents, and unchecking it removes all 26 rules at once.

The path is the same one Squarespace's help center documents²: "Open the Settings panel. Click Crawlers. Check the box next to 'Block known artificial intelligence crawlers.'" Read that as the opposite for owners who want citations — leave it unchecked, or uncheck it if a previous designer turned it on. There is no save confirmation pulse; the change writes to your live robots.txt within seconds.

One subtlety: the Squarespace UI state can briefly lag the live robots.txt file after a save. If the checkbox shows the right state but the live file does not, hard-refresh the panel, toggle once and save, then re-verify. The robots.txt itself is the authoritative source for what crawlers see.

 # If your robots.txt contains lines like these, GPTBot is being told to stay out User-agent: GPTBot Disallow: / User-agent: ClaudeBot Disallow: / # ...and 24 more Disallow blocks for the rest of the Squarespace AI list 

When the AI box is unchecked and GPTBot is allowed, none of those Disallow rules appear in your robots.txt. The file will still contain rules for non-AI bots (the standard User-agent: * block, anything Squarespace manages for sitemap and admin paths) — but GPTBot will be absent, which means the rules-default permits crawling.

 # Test with the published GPTBot user-agent string
curl -I -A "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.3; +https://openai.com/gptbot" https://yoursite.com/ # Expected: HTTP/2 200 OK (with normal response headers) # Problem state: HTTP/2 403 Forbidden, or anything in the 4xx/5xx range 

The user-agent string is published on OpenAI's bots documentation¹. OpenAI also publishes the IP range list at openai.com/gptbot.json⁵, which is useful if you want to cross-reference your server logs to see which GPTBot visits actually came from OpenAI versus user-agent-spoofers. The IP-based check is more thorough than the user-agent check, but for most Squarespace owners the curl one-liner is enough — Squarespace does not block user-agents at the platform level, so a 200 OK in this test means the bot can crawl.

The recovery curve has two parts. The fast part is OAI-SearchBot and the ChatGPT Search index — once the robots.txt is clean, OAI-SearchBot will re-fetch on its normal schedule and your URL will start surfacing in source cards again. The slow part is GPTBot's training contribution, which is a one-way function: anything not crawled during the training window is not in the model, and the next chance to be included is the next model version.

The honest framing for owners who toggled the box on years ago: you may have missed a training cycle, but you have not lost long-term citation potential. Squarespace's exclusion explicitly does not delete what was already scraped², and ChatGPT Search uses a live-retrieval crawler (OAI-SearchBot) that does not depend on training data to surface your page. Unblock the bot, fix the citation hygiene, and the surface comes back.