> ## Documentation Index
> Fetch the complete documentation index at: https://docs.peec.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# DeepSeek shared chat links are indexed by Google

> David Konitzny's analysis (Jul 2026) — DeepSeek's /share/ conversation URLs are Google-indexed; only 13.2% of observed queries triggered web search, and each triggered search generates five fan-outs.

DeepSeek's shared conversation links are indexed by Google and discoverable with a simple `site:` search on the /share/ directory. David Konitzny documented the mechanism ([full article](https://www.linkedin.com/pulse/when-ai-conversations-become-public-analysis-deepseek-david-konitzny-xn8ze/), Jul 2026): DeepSeek's share button warns that "anyone with the link can view" but never mentions the link can be indexed by search engines — repeating the ChatGPT incidents of 2024.

Usage findings from the indexed shared chats (small dataset, directional only, Jul 2026):

* Russian is the top language at 31% of observed shared links; English second at 20%.
* **Only 13.2% of observed queries triggered a web search** — most DeepSeek conversations run without retrieval. When search does trigger, **the system generates five fan-outs immediately**, dominated by year modifiers (2025/2026) and `site:` operators — the same patterns seen in other engines.
* Most users treat DeepSeek as a single-question tool; conversations mostly end before \~10 message exchanges.

Shared-link exposure is a web-wide pattern, not an AI-specific one, and those links can even enter AI retrieval pipelines — see [Shared URLs in AI retrieval](/research/shared-urls-in-ai-retrieval).

**Source:** [When AI conversations become public: an analysis of DeepSeek shared links](https://www.linkedin.com/pulse/when-ai-conversations-become-public-analysis-deepseek-david-konitzny-xn8ze/), David Konitzny, LinkedIn, Jul 2026.