Keenable SELECT：用SQL语言查询网络的AI Agent工具

Keenable推出了SELECT，一款创新型AI Agent工具，能够使用SQL语言对网络数据进行查询和语义提取 1。该工具通过MCP服务器运行，将网络搜索与语义操作集成到SQL查询框架中 1。

SELECT的核心功能包括select、WEB_SEARCH、WEB_FETCH和generate_html_report等操作 1。用户可在单次调用中搜索超过1000个页面，并通过精确的WHERE子句和语义操作符对结果进行过滤和字段提取 1。工具支持SEM_EXTRACT和SEM_MATCH等语义操作符执行内容提取与匹配 1，同时允许按行执行WEB_SEARCH和WEB_FETCH操作，其中参数可引用行列数据 1。该系统支持读取DuckDB SELECT语句，为用户提供灵活的网络数据查询体验 1。

Keenable has introduced SELECT, an innovative AI agent tool that enables users to query and extract semantic information from web data using SQL language 1. The system operates through an MCP server architecture, integrating web search and semantic operations directly into SQL queries 1.

The platform can search over 1,000 pages in a single call and filter results through precise WHERE clauses combined with large language model operators 1. Core functionality includes SELECT, WEB_SEARCH, WEB_FETCH, and generate_html_report features 1. Users can execute WEB_SEARCH and WEB_FETCH operations row by row, with parameters referencing column data from previous results 1. The tool employs semantic operators such as SEM_EXTRACT and SEM_MATCH to perform content extraction and pattern matching 1. All operations process DuckDB SELECT statements through the MCP server backend 1.