dbt Semantic Layer MCP Server

Overview

The dbt Semantic Layer MCP Server is a purpose‑built bridge that lets AI assistants—such as Claude Desktop—talk directly to the dbt Semantic Layer. By exposing a set of well‑defined resources, tools, and prompts over the Model Context Protocol, it turns complex SQL logic into conversational queries. Developers no longer need to hand‑craft analytics code; instead they can ask natural‑language questions and receive precise, metric‑driven answers that respect the single source of truth established in dbt.

Problem Solved

In many data environments, business metrics are scattered across spreadsheets, dashboards, and ad‑hoc queries. Even when a Semantic Layer exists, teams often struggle to access it from conversational AI tools because the layer is not exposed as an API. This MCP server eliminates that friction by providing a standardized, AI‑friendly interface to the Semantic Layer, enabling instant, consistent metric retrieval without exposing raw database credentials or SQL.

What It Does

• Metric discovery: The server lists all available metrics, allowing users to browse or search by name.
• Natural‑language query generation: Users can request metrics in plain English (e.g., “Show me monthly revenue by product category”) and the server translates that into a Semantic Layer query.
• Data filtering, grouping, and ordering: The MCP supports dynamic filters (date ranges, dimensions) and aggregation controls so analysts can drill down or roll up as needed.
• Result visualization: Query results are formatted into tables or charts that appear directly within the AI assistant interface, making insights immediately actionable.

Key Features

• 🔍 Metric Discovery – Browse or search metrics in the Semantic Layer.
• 📊 Query Creation – Convert natural‑language prompts into executable semantic queries.
• 🧮 Data Analysis – Apply filters, groupings, and sorting to refine insights.
• 📈 Result Visualization – Display results in clear, readable formats inside the assistant UI.

Use Cases

• Business analysts ask for up‑to‑date KPI reports without writing SQL.
• Data engineers expose new metrics to the team through a single conversational channel.
• Product managers explore trend data (e.g., week‑over‑week growth) in real time during meetings.
• Executive dashboards can be queried on demand via chat, reducing the need for separate BI tools.

Integration with AI Workflows

Once installed on Smithery, the MCP registers itself as a client‑side tool for Claude. In practice, developers simply configure the assistant to use this server; thereafter any conversation can invoke metric queries by referencing the server’s prompts. The assistant handles authentication, query translation, and result rendering automatically, letting developers focus on higher‑level business logic.

Standout Advantages

• Single source of truth: All metrics come from the dbt Semantic Layer, ensuring consistency across teams.
• Zero‑code interaction: Users interact purely through natural language, eliminating the barrier of SQL or API knowledge.
• Rapid prototyping: New metrics added to dbt become instantly available in the assistant without redeploying code.
• Built on Smithery: The server leverages a managed MCP platform, simplifying deployment and scaling.

This MCP server empowers developers and analysts to harness the full power of dbt’s Semantic Layer through conversational AI, streamlining data access and accelerating insight delivery.