# dbt [Health: Active]

**Category:** 📊 Data Platforms  
**Repository:** https://github.com/us-all/dbt-mcp-server  
**GitHub Stars:** 0  
**npm Downloads (last month):** 180  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/dbt

## Description
dbt MCP — manifest/run_results/sources/catalog parsing + DQ result tables (BigQuery/Postgres)

## Tools
Capabilities this server exposes over MCP:

- **dbt-list-models**
- **dbt-get-model**
- **dbt-list-tests**
- **dbt-get-test**
- **dbt-list-sources**
- **dbt-get-source**
- **dbt-list-exposures**
- **dbt-list-macros**
- **dbt-get-macro**
- **dbt-list-runs**
- **dbt-get-run-results**
- **dbt-failed-tests**
- **dbt-slow-models**
- **dbt-coverage**
- **dbt-graph**
- **freshness-status**
- **incident-context**
- **dbt-sla-status**
- **dq-list-checks**
- **dq-get-check-history**
- **dq-failed-checks-by-dataset**
- **dq-score-trend**
- **dq-tier-status**
- **dq-tier-by-source**
- **failed-tests-summary**
- **dq-score-snapshot**

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "dbt": {
    "command": "npx",
    "args": ["-y","@us-all/dbt-mcp"],
    "env": {
      "DBT_PROJECT_DIR": ""
    }
  }
}
```

**Requires environment variables:** `DBT_PROJECT_DIR` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What dbt MCP server does

The dbt MCP server provides read-only access to the information produced by a dbt project. It reads `manifest.json`, `run_results.json`, `sources.json`, and `catalog.json` from the project’s target directory, then exposes structured inspection tools through MCP. The available data supports model and test discovery, model and test details, source and exposure inspection, macro listing, run-history review, failed-test analysis, slow-model analysis, test coverage, graph traversal, source freshness, incident context, and SLA checks.

A separate quality category can query custom data-quality result tables in BigQuery or Postgres. Those tools cover check history, failed checks by dataset, score trends, tier status, and per-source tier results. The server also includes four prompts for failed-test investigation, freshness triage, DQ trend reporting, and incident triage.

## How it works

The server runs over MCP stdio by default. An MCP client starts the process with the dbt project directory supplied through the environment, and the server reads the local artifact files. `DBT_TARGET_DIR` can select a target directory other than the default under the project root. Archived run-results directories can also be supplied for historical analysis.

Quality queries use BigQuery by default or Postgres when `DQ_BACKEND=postgres`. The relevant result and score table names are provided through environment variables. Two schema presets are available: `generic` and `us-all`; individual column mappings can be adjusted with `DQ_COL_*` variables. Optional YAML configuration supplies dbt test and freshness thresholds plus tier targets.

The server includes aggregation tools that combine common multi-step lookups, and high-volume reads support response projection through `extractFields`. Tool categories can be enabled with `DBT_TOOLS` or disabled with `DBT_DISABLE`. Streamable HTTP is available by setting `MCP_TRANSPORT=http`; that mode includes Bearer authentication and a `/health` endpoint.

## Setup and configuration

Install the dbt MCP server as a development dependency with `pnpm add -D @us-all/dbt-mcp`, then run it with `npx @us-all/dbt-mcp`. Set `DBT_PROJECT_DIR` to the directory containing `dbt_project.yml`. Install `@google-cloud/bigquery` or `pg` only when using the corresponding custom DQ backend.

For BigQuery, authentication can use `GOOGLE_APPLICATION_CREDENTIALS`, with ADC fallback supported, and `BQ_PROJECT_ID` can explicitly select a project. Postgres uses `PG_CONNECTION_STRING`. `DQ_RESULTS_TABLE` is needed for checks-based quality tools, while `DQ_SCORE_TABLE` is needed for score-only tools. `DBT_ALLOW_WRITE` is reserved for future write tools; current tools remain read-only.

## Tools and capabilities

The `dbt` category contains artifact and analysis tools such as model, test, source, exposure, macro, and run access, along with failed-test, slow-model, coverage, graph, freshness, incident, and SLA analysis. The `quality` category reads DQ checks and score tables. The always-on `meta` category provides `search-tools` for natural-language tool discovery.

For DAG triggering, run orchestration, or log tails, use the companion Airflow MCP server instead. This server does not invoke `dbt run` and does not provide write operations.

_Full upstream README: https://allmcps.com/mcp/dbt/readme_