Claude dbt connectorWhat Claude can do in your dbt account.

Turns a question in plain English into SQL, using your project's context so table and column names match.

Lists the distinct values of a dimension, optionally limited to certain metrics, so Claude knows what it can filter on.

Returns the dimensions available for the metrics you name, the ways a number can be sliced.

Brings back the entities tied to the metrics you specify, which is what links a number to customers, orders or accounts.

Shows the compiled SQL behind a metric query without running it, so you can see exactly what the Semantic Layer would run.

Retrieves every metric defined in your Semantic Layer, the governed catalog of what can be queried.

Lists all saved queries in the Semantic Layer, the reusable requests your team already agreed on.

Runs a metric query with the filters and grouping you ask for, and returns the result for Claude to read out.

Retrieves the macros in your project, with an option to filter by package or return only package names.

Returns the name and description of every model in the project, a quick table of contents of the project.

Lists your sources together with their freshness status, optionally filtered by source name.

dbt's README marks this one deprecated and points you to a broader node-details tool that the directory sheet does not list. The sheet still carries the old name, so it gets a note here, without a current description of its own.

Gets all exposures, the downstream dashboards, apps or analyses that depend on your models.

Returns a bounded lineage graph around a resource, filtered by type, depth and direction.

Still on the directory sheet, yet flagged deprecated in dbt's README. The documentation names a general node-details tool as the replacement, and that replacement is not among the 60 tools the sheet lists.

Retrieves all mart models of the project, as a list of its own.

Deprecated according to dbt's README, which redirects to get_lineage, listed on this page with its type, depth and direction filters. The old name survives on the directory sheet only.

The README gives no live description for this name any more: it is deprecated, with a general node-details tool as the stated successor, one the directory sheet does not show.

Gathers health signals for a model: run status, test results and the freshness of its upstream sources.

Like the entry for children, this one is deprecated in dbt's README in favor of get_lineage, the bounded lineage tool listed further up. Only the directory sheet keeps the old entry.

Fetches a model's execution history, with an option to include test results alongside.

Finds models similar to the one you name, using semantic search rather than exact names.

Another entry the README labels deprecated. dbt names a general node-details tool as the one to use instead, though it is absent from the directory sheet, which still lists this legacy name.

Deprecated per dbt's README, with a general node-details tool given as the replacement. Since the directory sheet does not list that replacement, this page can only report the status of the old name, not describe what it once returned.

Flagged deprecated in the README, which points to a general node-details tool instead. The directory sheet keeps listing it; the documentation no longer explains it beyond that flag.

dbt's README lists it as deprecated and names a general node-details tool as the way forward. The old name stays on the directory sheet, while source freshness remains covered by get_all_sources, which is not deprecated.

The last deprecated detail tool in dbt's README, again with a general node-details tool as the named successor, absent from the directory sheet. The sheet's list keeps the old name regardless.

Searches for resources across the dbt project. dbt labels it alpha and says it is not generally available.

Turns models, tests and analyses into executable SQL, handy for checking Jinja logic before anything runs.

Reads lineage from the local manifest file, with type and depth filters, rather than from the dbt platform.

Pulls details for models, seeds, snapshots or sources from the local manifest file instead of the platform.

Lists resources in the project by type, with dbt's selector syntax to narrow the set.

Parses and validates the project files for syntax correctness, a quick sanity check on the whole project.

Runs your data tests to validate data and model integrity, and reports which ones pass or fail.

Gets a job's configuration: its triggers, its schedule and the dbt commands it runs.

Retrieves the details of one job run: status, timing, steps and artifacts.

Pulls the error details of a job run, with an option to include warnings or show warnings only.

Lists the artifacts available from a job run, such as the files dbt produces after it finishes.

Lists the jobs in your dbt platform account, optionally filtered by project or environment.

Lists job runs, with filters by job or status and ordering by a field of your choice.

Lists all projects in your dbt platform account, as far as the account you connected can see them.

Compiles SQL in the context of your project, through the dbt platform rather than locally.

Traces lineage at the column level through the dbt platform, finer than model-level lineage.

Traces column-level lineage locally, which requires dbt's language server from the dbt Labs VS Code extension.

Fetches the full Markdown content of one or more pages from dbt's public documentation at docs.getdbt.com, by path or URL.

Searches dbt's public documentation and returns matching page titles, URLs and descriptions ranked by relevance.

Returns the current git branch of the running dbt server, a debugging detail more than a daily tool.

Reports the current version of the dbt server Claude is connected to.