// MCP TOOL REFERENCE

DBT TOOLS

55 tools from the Dbt MCP Server, categorised by risk level.

// ALL 55 DBT TOOLS

READ 46 tools

Read compile Generate executable SQL from models without running Read docs Generate documentation for the dbt project Read get_all_macros Retrieve macros with optional package filtering Read get_all_models Retrieve name and description of all models Read get_all_sources Get all sources with freshness status Read get_column_lineage Trace column-level lineage locally Read get_dimensions Get dimensions for specified metrics Read get_entities Get entities for specified metrics Read get_exposure_details Get exposure details including owner and freshness Read get_exposures Get all exposures Read get_job_details Get job config including triggers and schedule Read get_job_run_artifact Download a specific artifact from a job run Read get_job_run_details Get run details including status and timing Read get_job_run_error Get error and warning details for a job run Read get_lineage Get full lineage graph with type and depth filtering Read get_lineage_dev Retrieve lineage from local manifest.json Read get_macro_details Get details for a specific macro Read get_mart_models Retrieve all mart models Read get_mcp_server_branch Return the git branch of the MCP server Read get_mcp_server_version Return the version of the dbt MCP server Read get_metrics_compiled_sql Get compiled SQL for metrics without executing Read get_model_children Get downstream dependents of a model Read get_model_details Get model details including compiled SQL and columns Read get_model_health Get health signals for a model Read get_model_parents Get upstream dependencies of a model Read get_model_performance Get execution history for a model Read get_node_details_dev Retrieve node details from local manifest.json Read get_project_details Get project information Read get_related_models Find similar models using semantic search Read get_seed_details Get details for a specific seed Read get_semantic_model_details Get details for a specific semantic model Read get_snapshot_details Get details for a specific snapshot Read get_source_details Get source details including columns and freshness Read get_test_details Get details for a specific test Read list List resources in the dbt project by type Read list_job_run_artifacts List available artifacts from a job run Read list_jobs List jobs in a dbt Cloud account Read list_jobs_runs List job runs with optional filtering Read list_metrics Retrieve all defined metrics Read list_saved_queries Retrieve all saved queries Read parse Parse and validate project files Read query_metrics Execute metric queries with filtering and grouping Read search Search for resources across the dbt project Read show Execute SQL and return results without persisting Read test Run tests to validate data and model integrity Read text_to_sql Generate SQL from natural language without executing

WRITE 3 tools

Write generate_model_yaml Generate model YAML with column definitions Write generate_source Generate source YAML by introspecting schemas Write generate_staging_model Generate staging model SQL from a source table

DESTRUCTIVE 1 tools

Destructive cancel_job_run Cancel a running job

EXECUTE 5 tools

Execute build Execute models, tests, snapshots in DAG order Execute execute_sql Execute SQL on dbt Cloud infrastructure Execute retry_job_run Retry a failed job run Execute run Execute models to materialise them in the database Execute trigger_job_run Trigger a job run on dbt Cloud

// RUNNING THIS SERVER

The managed route: connect Dbt through the PolicyLayer gateway — every tool call above is checked against your policy before it runs, with a full audit log.

DIRECT INSTALL (UNMANAGED) npx -y @dbt-labs/dbt-mcp

// FAQ

How many tools does the Dbt MCP server have? +

The Dbt MCP server exposes 55 tools across 4 categories: Read, Write, Destructive, Execute.

How do I enforce policies on Dbt tools? +

Route the Dbt server through the PolicyLayer gateway. Define allow, deny, or approval rules per tool in the dashboard — they are enforced on every call before it reaches the server.

What risk categories do Dbt tools fall into? +

Dbt tools are categorised as Read (46), Write (3), Destructive (1), Execute (5). Each category has a recommended default policy.

Let agents act without letting them run wild.

Route your MCP servers through PolicyLayer and every tool call is checked against your policy before it runs — allow, deny, or require approval. Per-identity grants. Full audit log. Live in minutes.

SECURE YOUR MCP →

Free to start. No card required.

4,600+ MCP servers and 31,000+ tools scanned and risk-classified.