Documentation gap

The Azure MCP Server Azure AI Search documentation does not explain how search_index_query selects its search behavior. This caused #2297 to report hybrid and vector support as missing even though the implementation already constructs a hybrid request when the index is configured for vectorization.

Scope

Update the Azure MCP Server Azure AI Search documentation, specifically the Query search index section, to explain:

• An index without vector fields uses keyword search.
• When vector fields have a configured vectorizer, the tool sends the query text as both a lexical query and vector queries, and Azure AI Search combines the results using Reciprocal Rank Fusion (RRF).
• Hybrid search requires vector fields, vector search profiles, and a working vectorizer configuration on the index.
• The tool does not expose vector-only search, semantic ranking, multimodal search, or explicit query-type controls.
• One example prompt for querying a vectorizer-enabled index and the expected hybrid behavior.

Keep the wording precise: the current implementation always includes lexical search text, so it should not claim that the tool automatically performs vector-only search.

Related work

• Customer report and confirmed documentation gap: [Feature Request] Add hybrid/vector search support to Query Index tool #2297
• Prior incomplete-description issue: Azure MCP for AI Search index query - tool description incomplete #1903
• Prior tool-description PR, closed because the change belongs in public documentation: AI Search index query tool description improvement #1904
• Broader query configuration request, including explicit query type and Full query syntax: MCP Performs only basic types of query to Azure Search (azuremcp_search_index_query) #219

The source page is MicrosoftDocs/azure-dev-docs/articles/azure-mcp-server/tools/azure-ai-search.md.