Log inSubmit server
Dbt MCP logo

Dbt MCP

by dbt-labs

Connects AI agents to dbt projects, Semantic Layer metrics, Discovery metadata, Admin API operations, and dbt documentation.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget

Install Dbt MCP

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "dbt-labs-dbt-mcp": {
      "command": "uvx",
      "args": [
        "dbt-mcp"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— the server itself is confirmed live (we called its tools/list endpoint). The stdio install command is a separate, ongoing check.

Overview

dbt-labs/dbt-mcp connects AI agents with dbt Core, dbt Fusion, and dbt Platform through tools for CLI execution, Semantic Layer queries, project discovery, and Admin API operations. It can inspect models, sources, exposures, macros, lineage, health, performance, and project documentation, while also generating dbt code and SQL. Semantic Layer tools support metric discovery, dimension and entity inspection, compiled SQL generation, and filtered queries. Reach for it when an agent needs to answer questions about a dbt environment, run trusted dbt operations, investigate model dependencies, or query governed business metrics.

Use cases

  • Query governed business metrics through the dbt Semantic Layer
  • Inspect model metadata, health, performance, and lineage
  • Run and validate dbt CLI commands from an agent
  • Investigate dbt Platform jobs and job run artifacts
  • Generate staging models and dbt YAML files

Key features

  • dbt CLI command execution
  • Semantic Layer metric discovery and querying
  • Model, source, macro, and exposure discovery
  • Bounded lineage and column-level lineage
  • dbt Platform Admin API operations
  • Product documentation search and retrieval

Tools39

What Dbt MCP lets your agent do. Expand a tool for its parameters.

Verified live~20.8k tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 1d ago

Recent health check succeeded.

14/14 checks healthy over the last 78d

What it can do

From the 39 tools this server returned when we ran it; 39 declare their own safety hints.

Runs commands · 3Handles credentials · 2Makes network requests · 39Read-only tools · 36
Unusual tool instructions. 7 tool descriptions contain text aimed at the AI model rather than at you (hidden instruction block). Read the tool descriptions before connecting.
Network at startup2 connections
Files written at startupNone

Observed 2026-10-18 in an isolated sandbox. A description of reach, not a security audit.

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 2Low 6

Scanned 11d ago via OSV.dev · dbt-mcp (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars613Stargazers on the source repository.
Last commit1d agoMost recent push to the default branch.
Tools exposed39Callable tools this server registers over MCP.
Directory activity10 viewsConfig copies, upvotes, and views on AllMCPs.

How Dbt MCP works

What dbt-labs/dbt-mcp does

The dbt-labs/dbt-mcp MCP server gives AI agents access to dbt Core, dbt Fusion, and dbt Platform capabilities. Its coverage spans local project work, dbt Platform infrastructure, metadata discovery, the Semantic Layer, administrative operations, code generation, and official dbt product documentation.

Agents can inspect models, mart models, sources, exposures, macros, seeds, snapshots, tests, and semantic models. They can also examine lineage, source freshness, model health, execution history, job configuration, job runs, and project information. Resource details are available through get_node_details, while older resource-specific detail tools are marked deprecated.

How it works

The server exposes MCP tools that an agent can call according to the task. Discovery tools return paginated results for collections such as models, sources, jobs, projects, exposures, and macros. Lineage queries return connected nodes and explicit parent-to-child edges, with controls for depth, direction, resource type, and result limits.

Semantic Layer workflows generally begin by listing metrics, then inspecting their dimensions or entities before calling query_metrics. Agents can retrieve valid dimension values before constructing filters and can request compiled SQL without executing a metric query. Saved queries are available for discovering predefined analytical requests.

Product documentation tools search and fetch Markdown from docs.getdbt.com. Search returns page metadata and can fall back to full-text matching; page content is retrieved separately with get_product_doc_pages. The server also reports its running branch and version.

Tools and capabilities

  • Execute, compile, parse, build, run, test, show, clone, list, and document dbt projects through CLI tools.
  • Query the dbt Semantic Layer, list metrics and saved queries, inspect dimensions and entities, retrieve dimension values, and generate compiled SQL.
  • Browse dbt resources and metadata, including models, mart models, sources, exposures, macros, and semantic models.
  • Trace upstream and downstream lineage, inspect model health and performance, and review source freshness.
  • Manage dbt Platform projects and jobs, including triggering, retrying, cancelling, and inspecting job runs and artifacts.
  • Generate model YAML, source YAML, and staging model SQL; compile SQL and inspect column lineage through dbt LSP-related tools.
  • Search and retrieve official dbt product documentation.

Limitations and notes

CLI tools can modify data models, sources, and warehouse objects, so they should be used only with a trusted client and an understanding of the requested command's effects. Semantic Layer queries require valid metrics, dimensions, and entities; the query tools are not intended for questions that cannot be expressed with the available semantic definitions.

Large collection and lineage responses are bounded or paginated. Lineage results report omitted nodes when a limit excludes part of the graph. Metric listings may omit lower-priority columns to stay within response limits; narrower searches can retrieve details for selected metrics.

Several older tools, including resource-specific detail tools and separate parent or child lineage tools, are deprecated in favor of get_node_details and get_lineage. The repository also publishes an experimental dbt-mcp.mcpb bundle for MCPB-aware clients.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Dbt MCP

dbt-labs/dbt-mcp is an MCP server that connects AI agents to dbt Core, dbt Fusion, and dbt Platform. Its main tools cover dbt CLI commands, Semantic Layer metric discovery and queries, Discovery metadata and lineage, Admin API operations, code generation, LSP compilation, and dbt product documentation.

Install the experimental MCP bundle by downloading dbt-mcp.mcpb from the latest release assets and following the Anthropic mcpb CLI documentation. MCPB-aware clients can import the bundle without additional setup.

Yes, Cursor is named as an example agent product that can connect to dbt-labs/dbt-mcp. The repository also describes connections from Claude-based agent products and supports building a custom agent.

Yes, it can list metrics and saved queries, inspect dimensions and entities, discover dimension values, execute metric queries, and return compiled SQL without executing the query. Queries support grouping, ordering, filtering, and time grains when the requested fields are available.

Yes, it provides lineage graphs with upstream or downstream direction, depth, type, and result-limit controls. Separate tools report model run status, test status, upstream freshness, and historical execution performance.

Related MCP servers

All in Data Platforms All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewDbt MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/dbt-labs-dbt-mcp?style=directory)](https://allmcps.com/mcp/dbt-labs-dbt-mcp)
HTML Embed
<a href="https://allmcps.com/mcp/dbt-labs-dbt-mcp"><img src="https://allmcps.com/api/badge/dbt-labs-dbt-mcp?style=directory" alt="Dbt MCP on AllMCPs" /></a>