Seed Phrases, Nine-Tailed Foxes, and a Missing HTML Tag: Claude Chats Go Google

Hundreds of conversations shared through Anthropic's Claude platform were indexed by Google after a single missing line of HTML code, exposing everything from a user's cryptocurrency wallet seed phrase to a lawyer's ethics question and a person documenting their progress toward becoming a nine-tailed fox. A Reddit user surfaced the issue on July 25 by running the query site:claude.ai/share, which returned hundreds of third-party links to "Anyone with the link" share pages that Google had indexed despite Anthropic's intent to keep them unlisted.

The exposure stemmed from the absence of a noindex meta tag, the HTML instruction telling search engines not to list a specific page. Anthropic's robots.txt file — the standard text file websites use to tell crawlers which pages to skip — was already blocking Claude's share URLs, but Google's own documentation notes that blocking a crawl does not prevent indexing when the URL appears elsewhere on the web. Because the crawler could not open the page, it also could not read the missing noindex instruction, so search results surfaced the links labeled "No information is available for this page." The same gap affected Claude's published artifacts; a search for site:claude.ai/public/artifacts surfaced payroll spreadsheets with employee names, internal customer relationship management chat exports, and unreleased product roadmaps.

The episode echoes a similar incident involving OpenAI in August 2024, when its "Make this chat discoverable" checkbox sent thousands of ChatGPT conversations to Google, Bing, and DuckDuckGo until OpenAI CISO Dane Stuckey removed the feature, calling it "a short-lived experiment to help people discover useful conversations." Many of those chats were captured by the Internet Archive, a nonprofit that permanently stores copies of webpages, before OpenAI could remove them and remain accessible today. Anthropic, which spent April lobbying against similar third-party indexing practices, has not yet disclosed the total number of affected Claude shares or whether any of the indexed data has been removed from Google's cache.

Users who shared conversations through Claude during the affected period were advised to revoke any published share links and treat any sensitive material — including seed phrases, internal documents, or unreleased product details — as compromised, since indexed copies can persist in web archives even after the original pages are deleted.