Which sources do AI models actually cite?
Track Reddit, Wikipedia, G2, LinkedIn, and 180+ domains across ChatGPT, Claude, Perplexity, Gemini, Grok, and DeepSeek.
Sources by AI model
OpenAI
OpenAI's conversational AI. The most widely used LLM, with browsing and retrieval capabilities that pull from a broad range of web sources.
Anthropic
Anthropic's AI assistant. Known for detailed, sourced responses with web search capabilities.
Perplexity AI
An answer engine built around citations. Every response includes numbered source links from live web retrieval.
Last seen: source activity across all models
When was each source last cited? 15 sources tracked across 3 AI models.
All tracked sources by category
💬 Forum Sources
Discussion forums and Q&A sites
reddit.com
The internet's largest forum. Subreddits cover every topic, from product recommendations to technical troubleshooting. LLMs frequently cite Reddit threads as first-hand user experiences.
stackoverflow.com
The largest developer Q&A platform. A top source for code examples, error resolution, and technical explanations in LLM responses.
quora.com
Q&A platform covering broad topics. Cited for expert opinions, personal experiences, and long-form answers.
news.ycombinator.com
Y Combinator's tech community. Cited for developer opinions, startup discussions, and tech industry commentary.
⭐ Review Sources
Software and product review sites
g2.com
The largest B2B software review platform. LLMs cite G2 reviews for software comparisons, ratings, and buyer sentiment.
sourceforge.net
Open-source software directory and review platform. Cited for software downloads, reviews, and project comparisons.
capterra.com
B2B software review and comparison platform. Cited for software category lists, feature comparisons, and user reviews.
trustpilot.com
Consumer review platform. Cited for business reputation, customer sentiment, and service quality assessments.
👥 Social Sources
Social networks
linkedin.com
The professional social network. LLMs cite LinkedIn for company profiles, professional bios, and industry discussions.
x.com
Real-time social platform. Grok has native access; other LLMs cite X posts for trending topics, opinions, and announcements.
youtube.com
The largest video platform. LLMs reference YouTube for tutorials, reviews, and educational content via metadata and transcripts.
📚 Wikipedia Sources
Wikipedia and wikis
📖 Documentation Sources
Technical docs and code repos
📋 Directory Sources
Product directories and listings
📰 Media Sources
News and publications
How to see what sources ChatGPT uses
ChatGPT, Claude, Perplexity, and other LLMs pull from different source pools. Some cite live web pages. Others draw from training data that includes Reddit, Wikipedia, and news sites.
The sources an LLM cites directly shape how it describes your brand, recommends your product, and positions you against competitors. If G2 reviews say you're the best CRM, ChatGPT will repeat that claim.
xSeek tracks citation patterns across all major AI models. You see exactly which domains get cited, how often, and whether your brand appears. That's the difference between hoping LLMs say the right thing and knowing they do.
Best tools to check ChatGPT sources
Native UIs like Perplexity and SearchGPT show inline citations. Browser dev tools reveal the actual URLs fetched during retrieval. Enterprise teams use observability layers that log every source for every query.
For ongoing monitoring, tracking whether your brand gets cited more or less over time, you need an AI visibility tool. xSeek monitors citation frequency, share of voice, and source patterns across ChatGPT, Claude, Perplexity, Gemini, Grok, and DeepSeek.
Frequently asked questions
How to see what sources ChatGPT uses?
You can check ChatGPT's sources by enabling browsing mode (SearchGPT) which shows inline citations, inspecting network requests in browser dev tools, or using an AI visibility tool like xSeek that tracks citation patterns across ChatGPT responses at scale.
Is Reddit still used to train LLMs?
Yes. OpenAI has a $60M/year licensing deal with Reddit. Reddit content appears in ChatGPT training data and in live browsing citations. Other LLMs also reference Reddit threads through web search retrieval, though citation frequency has shifted since mid-2025.
What are the best tools to check ChatGPT sources?
Native citation UIs (Perplexity, SearchGPT), browser network inspection for live URL retrieval, retrieval observability layers for custom LLM stacks, and AI visibility platforms like xSeek that monitor citation frequency and share of voice across all major LLMs.
How often does ChatGPT cite Reddit in 2025-2026?
Reddit URLs appear in approximately 8-15% of ChatGPT responses that include citations, depending on query type. Consumer queries and product recommendations cite Reddit most frequently. The rate has fluctuated but Reddit remains a significant source.
What is generative engine optimization (GEO)?
GEO is the practice of optimizing your content and online presence to be cited by AI answer engines like ChatGPT, Perplexity, and Claude. It involves monitoring which sources LLMs cite, ensuring your brand appears on high-citation domains, and structuring content for AI retrieval.
Do different LLMs use different sources?
Yes. Each LLM has distinct citation patterns. Perplexity cites the widest variety of sources in every response. Grok has unique access to X/Twitter. Gemini leverages Google Search. ChatGPT varies between training data and live browsing. Claude favors documentation and academic sources.