Perplexity and Claude omit websites primarily due to three technical barriers: edge firewall blocks (HTTP 403/503 on PerplexityBot or ClaudeBot), lack of a machine-readable /llms.txt directory, or thin client-side JavaScript pages that lack server-rendered text and JSON-LD schema.
Perplexity and Claude synthesize answers by retrieving and ranking clean, high-authority markdown or semantic HTML passages in real time. If retrieval fails or times out, competitors with cleaner grounding win the citation.
Verify that PerplexityBot and ClaudeBot receive HTTP 200 responses under 500ms, deploy a root /llms.txt manifest pointing to your core documentation, and embed WebSite/Organization JSON-LD schema.