orionbelt-semantic-layer

概述

OrionBelt 语义层是一种 API 优先的语义引擎，它将声明性 YAML 模型编译为针对多个数据库平台的优化、特定于方言的 SQL，包括 BigQuery、PostgreSQL、Snowflake、ClickHouse、Dremio、Databricks、DuckDB 和 MySQL。它使 AI 代理能够使用业务概念而非原始 SQL 查询数据库，通过 REST API 和 MCP 集成实现该解决方案。该版本提供了代码即分析功能，具有版本控制的 YAML 模型、特定于方言的 SQL 生成，以及包括 UI、API 和 MCP 服务器在内的完整集成界面。

试试问 AI

装完之后，这里有 6 个你可以让 AI 做的事：

• compile_query

  Compile a query into SQL based on the semantic model

  试试:Calculate total revenue by country试试:Show me the top 5 products by sales试试:What's the monthly trend of user registrations?
• validate_model

  Validate a semantic model definition

  试试:Check if my model definition is valid试试:Validate the sales model for errors试试:Is this OBML model correctly structured?
• list_models

  List available semantic models

  试试:Show me all available models试试:What models are loaded in the system?试试:List the marketing analytics models
• query_plan

  Get query plan information without executing the query

  试试:Show me the query plan for this request试试:How would you join these tables?试试:Explain the join strategy for multi-fact query
• find_semantics

  Find semantic elements in the model

  试试:Find all dimensions related to time试试:Search for revenue metrics试试:What measures are available for product data?
• execute_query

  Execute a compiled query against the database

  试试:Run this SQL query against BigQuery试试:Execute the customer analysis query试试:Get results for the monthly sales report
• compile_examples

  Get example queries from model definitions

  试试:Show me examples from the sales model试试:What queries are available in the marketing model?试试:Get example use cases for the financial model
• heartbeat

  Invalidate cache for dependent queries

  试试:Mark customer table as updated试试:Refresh cache for sales data试试:Invalidate all queries depending on orders
• get_model_health

  Check model health including orphaned elements

  试试:Check the health of the sales model试试:Are there any fan-traps in my model?试试:Report on model integrity issues
• export_model

  Export model in different formats

  试试:Export the model as RDF/OWL试试:Convert OBML to OSI format试试:Export the model schema as graph
• get_warnings

  Get warnings from model operations

  试试:Show me warnings from the last compilation试试:Are there any issues with my model?试试:Report on validation warnings
• discover_databases

  Discover available databases and schemas

  试试:What databases are available?试试:List all schemas in the data warehouse试试:Show me the customer data sources

可对比工具

安装

pip install orionbelt-semantic-layer-mcp

{
  "mcpServers": {
    "orionbelt": {
      "command": "uvx",
      "args": ["orionbelt-semantic-layer-mcp"]
    }
  }
}

pip install orionbelt-semantic-layer

docker run -p 8080:8080 ralforion/orionbelt-api

FAQ

OrionBelt 支持哪些数据库？

OrionBelt 支持 BigQuery、PostgreSQL、Snowflake、ClickHouse、Dremio、Databricks、DuckDB/MotherDuck 和 MySQL。

OrionBelt 如何处理 SQL 注入防护？

OrionBelt 使用基于 AST 的 SQL 生成，而非字符串模板，这确保了在所有支持的方言中生成防注入的 SQL。

我能否将 OrionBelt 与 Claude 以外的 AI 助手一起使用？

是的，MCP 服务器也适用于 Copilot、Cursor 和 Windsurf。主 API 可以与任何支持 REST 或 HTTP 端口的平台集成。

orionbelt-semantic-layer 对比