semantic_search_ab_tests

AI-powered semantic search using pgvector embeddings. Understands meaning, not just keywords.

Parameters

Parameter	Type	Required	Description
query	string	yes	Natural language query
max_results	integer	no	1–50, default 15

When to use this

Use semantic search for natural language questions where keywords alone won't work:

• "How do I make users trust a SaaS pricing page?"
• "What makes hero sections convert better?"
• "Best practices for reducing checkout abandonment"

For specific keyword searches (e.g. "CTA button red green"), use search_ab_tests instead.

Example prompts

"How do I reduce friction in a checkout flow?"

"What kind of social proof works best for B2B SaaS?"

"Should I use long or short hero headlines?"

Output format

Same as search_ab_tests, with similarity scores instead of keyword match percentages.