Google Indexed Claude Shared Chats via Search
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Reddit users in a Claude-related forum revealed that the search query
site:claude.ai/sharepulled up shared chats spanning adult fanfiction, fake book reviews, news discussions, and programming conversations. - Anthropic appeared to act quickly after being notified — the Google dork stopped returning results soon after the Reddit thread surfaced, per the thread itself.
- Private Claude chats were not exposed; only conversations where users had intentionally clicked the Share button and generated a link were indexed by Google.
- ZDNET's review of the still-active shared links found deeply personal content with potential financial or legal repercussions, including unverified claims about chats involving cryptocurrency wallets.
- The incident mirrors similar 2025 leaks of ChatGPT and Grok shared conversations pulled up in Google results, suggesting a recurring gap in how AI chatbot share features handle search-engine crawling.
- Anthropic has not yet publicly responded to ZDNET's request for comment on whether it will disable affected links server-side rather than relying on individual users to clean up their own share history.
- Claude's share feature creates a public snapshot accessible via URL but ships with no apparent server-side guardrail — such as a
noindextag — to stop Google and other engines from mass-indexing shared links.
Why it matters: Any user who clicked Claude's Share button believing only their intended recipient would see the conversation now has a publicly indexed link they may not know about, and Anthropic has not committed to scrubbing those URLs — meaning affected users must audit Settings > Privacy > Shared Chats and disable active links one-by-one to stop further exposure.
