dbt-doctor

Active·★ 138·MIT·Updated 2026-05-02

★ Trending★ RAG / Knowledge Base★ Security & Safety

AI-driven quality & governance MCP Server for dbt projects. Audit coverage, profile data, detect schema drift, and auto-generate documentation — all through natural language with your AI assistant.

dbt-doctor is an MCP server that gives AI coding assistants deep awareness of dbt project health. It enables natural language auditing, profiling, schema drift detection, and auto-documentation without manual CLI commands. It complements the official dbt-labs/dbt-mcp by focusing on analysis and generation rather than execution.

#audit#dbt-packages#mcp-server#mcp-servers

Features

01Project Auditing: score project 0–100% on documentation, testing, and naming conventions with ranked worst models.

02Data Profiling: single-pass column statistics (NULL rate, cardinality, min/max, uniqueness) using batched SQL queries.

03Schema Drift Detection: compare current warehouse columns against manifest.json to detect added, removed, or type-changed columns.

04Intelligent Test Suggestions: convert profile statistics into concrete dbt test recommendations (e.g., not_null, unique, accepted_values).

05Non-Destructive YAML Writing: update schema.yml files preserving hand-written comments and formatting using ruamel.yaml.

Compatibility

PostgreSQL

Postgres

Verified via docs

DuckDB

Verified via docs

Quick start

$ pip install dbt-doctor

Use cases

↳Auto-document the worst covered model in a project with a single AI request.

↳Detect schema drift before deployment by comparing warehouse columns to the manifest.

↳Perform DAG health checks to identify orphans, high fan-out, and structural issues.

Alternatives

mindsdb★ 39.2k

Federated Query Engine for AI - The only MCP Server you'll ever need

Features

Compatibility

Quick start

Use cases

Alternatives

Related searches

Comments

Features

Compatibility

Quick start

Use cases

Alternatives

Related searches

Comments