Wren Engine 是用于 AI 代理的开放上下文引擎，能够将业务数据转化为代理可理解的上下文。
Wren Engine 是一个用于 AI 代理的开放上下文引擎，能将业务数据转化为代理可用的上下文。它捕获业务定义、指标、关系和访问规则，使代理能够理解模型并生成准确的查询。该引擎支持 MCP、HTTP API，并可嵌入到各种代理工作流中。
Finally, a semantic layer that actually works for LLMs without requiring us to rewrite our entire database schema.
Connected this to our PostgreSQL and BigQuery instances. It translates natural language questions into highly accurate semantic queries.
Built on DataFusion, so the query planning is incredibly fast. We paired it with a Docker-based deployment in just a few minutes.
Does this support dynamic schema updates or do we need to restart the engine every time a table changes?
