Keenable SELECT: An AI Agent That Searches the Web Using SQL
Keenable AI introduced SELECT, a Model Context Protocol (MCP) agent designed to execute web searches and deep research using SQL queries paired with semantic operators. Built on Keenable's independent web search infrastructure, the system allows agents to retrieve and analyze live web data programmatically. By applying structured SQL syntax to web search, Keenable SELECT provides LLM agents with a precise and deterministic method for filtering unstructured online information. This approach simplifies the creation of automated deep-research workflows compared to traditional unstructured natural language search prompts. SELECT operates as an MCP agent that runs directly against Keenable's proprietary web crawler, search index, and ranking system. To demonstrate its capability, Keenable published 20 automated research reports generated entirely by SELECT querying real-time web data.
## BACKGROUND
SQL (Structured Query Language) is the standard language used to manipulate and retrieve data from relational databases. Model Context Protocol (MCP) is an open standard that allows AI models to seamlessly connect to external tools, API endpoints, and data repositories.