A MCP (Model Context Protocol) server for interacting with dbt resources. Current version: 1.16.0. Released 2026-04-27. Follows semantic versioning with frequent releases (biweekly).
pip install dbt-mcpVerified import paths — ran on the pinned version, not inferred.
Start a basic dbt MCP server. Requires dbt Cloud account ID and token set in environment.
Use `list_artifacts` and `get_artifact` (if available) or custom API calls.
Migrate to multi-project server config with `config_override` or use `list_projects`.
Remove `DBT_LSP_PATH`; use default Fusion LSP or editor storage paths.
Ensure Python 3.12 or later.
Upgrade to >=1.15.0 or avoid symlinked paths for `~/.dbt/mcp.lock`.
Update metric parsing to handle CSV (or update to use new threshold `DBT_MCP_SL_FULL_CONFIG_THRESHOLD`).
No dependency data recorded yet.