SIDE BY SIDE, read 10 Nov 2026
Cube MCP server and dbt MCP Server, field by field: 30 and 52 tools named
| Field | Cube MCP server | dbt MCP Server |
|---|---|---|
| Data tool | Cube | dbt |
| Publisher | Cube | dbt Labs |
| Kind | Official | Official |
| Repository or docs | docs.cube.dev/docs/integrations/mcp-server | github.com/dbt-labs/dbt-mcp |
| Version or commit | Hosted service (no versioned release) | 2.10.1 |
| Last release | none found on 10 Nov 2026 | 3 Nov 2026 |
| Transport | Remote (hosted) | Remote and local |
| Auth | OAuth Authorization Code + PKCE (scope mcp-agent-access, public client cube-mcp-client), or a Cube API key exchanged for a token (client_id cube-api-key). | dbt service/personal token and environment IDs for platform tools; OAuth or token per docs; local dbt project for CLI tools. |
| Access | Read and write | Configurable |
| Tools listed | 30 named in the evidence | 52 named in the evidence |
| Tool definitions read at | docs.cube.dev/docs/integrations/mcp-server.md | raw.githubusercontent.com/dbt-labs/dbt-mcp/main/README.md |
| Context weight | 1,093 tokens for 30 tool definitions (30 with a description), measured from the docs page read | 1,061 tokens for 63 tool definitions (63 with a description), measured from the README read |
| Status | Exists on 10 Nov 2026 as a Cube-hosted endpoint cubecloud.dev/mcp (or your own console domain). The page says it exposes 30 tools; the tables list 28 names, all captured here. | Exists, v2.10.1 on 3 Nov 2026 (PyPI and GitHub). |
| Read | 10 Nov 2026 | 10 Nov 2026 |
In one sentence each
- Cube MCP server: Cube's hosted server searches the semantic layer, runs SQL with row-level security as the signed-in user and edits the semantic model on a branch.
- dbt MCP Server: dbt Labs' server combines the dbt CLI, the Discovery API with model and column lineage, the Semantic Layer and SQL tools on dbt Platform.
The read facts, topic by topic
| Topic | Cube MCP server | dbt MCP Server |
|---|---|---|
| Publisher and releases | Cube hosts the MCP server and documents it on docs.cube.dev, available on all plans. No versioned release is published; one endpoint cubecloud.dev/mcp serves every account, deployment and region. | dbt Labs publishes dbt-mcp; the latest GitHub release is v2.10.1, whose release timestamp on the page is 3 Nov 2026 (PyPI also 3 Nov 2026). dbt Labs documents the MCP server on docs.getdbt.com and ships it as self-hosted or remote. |
| Clients, auth and scoping | Every tool runs as the authenticated user, and queries respect Cube permissions including row-level security. With an API key, a token can act as the key's owner, as another user, or be limited to specific tools. Which tools a user sees depends on their role; Viewer role or higher is needed, and data model editing tools show only for roles allowed to edit the model (Admin and Developer by default). | The dbt MCP server provides AI agents with context of a project in dbt Core, dbt Fusion and dbt Platform. |
| Access and safety | Nine tools are annotated destructive and clients that honor annotations always ask before running them. Chat is annotated read-only but runs the Analytics Chat agent, which can create reports or delete folders without confirmation, so the annotation understates what it can do. File writes never touch production: writeDataModelFile and deleteDataModelFile reject any branch that isn't a dev branch. API-key tokens last up to one hour and there is no refresh token. | dbt CLI tools can modify data models and warehouse objects, so the README warns to proceed only if you trust the client. dbt docs say the server stores no production data and is a read-only access layer reading from dbt platform in real time. Configuration uses a dbt service token (DBT_TOKEN) and tool groups can be enabled or disabled by environment variables. |
| What it reaches | Documented clients are Claude, ChatGPT, Microsoft 365 Copilot, Cursor and Codex, plus other MCP clients over HTTPS. Admins choose the default deployment and can limit which deployments MCP clients can reach, always intersected with role-based permissions. | Discovery tools cover models, sources, exposures (downstream dashboards, apps) and freshness. |
| Lineage and impact | SearchDataModel searches the semantic model by similarity, covering views and their measures and dimensions, and returns name, title, description and type. ListDataModelFiles and readDataModelFile read the raw YAML, JavaScript or Python source of the semantic model, so the definition logic itself is readable. None found: A lineage tool in the Cube MCP tool tables (getBranchDiff and getDataModelChanges compare model versions, not data lineage). | Get_lineage returns a bounded lineage graph with type, depth and direction filters. Column-level lineage tools exist: fusion.get_column_lineage via dbt Platform and get_column_lineage locally. Get_lineage_dev reads lineage from the local manifest.json. |
| Definitions, ownership and rules | Model editing tools (startDataModelEdit, writeDataModelFile, commitDataModelChanges) let an agent change definitions on a personal dev branch, merged to the default branch only with a confirm-prompting tool. Query results carry the semantic layer: runQuery runs Cube SQL against views whose names and members come from searchDataModel. | Semantic Layer tools list metrics, dimensions, entities and saved queries. Get_all_models returns model names and descriptions; generate_model_yaml can inherit upstream descriptions. |
| Freshness and upkeep | The server can target a dev-mode branch, so agents can query a branch's model before it is merged. GetPreAggregationStatus and buildPreAggregation let an agent check and trigger pre-aggregation builds. | Get_model_health returns run status, test results and upstream source freshness from dbt Platform. |
| Profile | Admins need admin privileges or a custom role with the Manage MCP server and connectors permission to enable MCP. | The README publishes an experimental MCP bundle (dbt-mcp.mcpb) with each release. |
Tool names, as published
Cube MCP server
listDeployments chat loadQueryResults searchDataModel runQuery readWorkbook createWorkbook createReport readReport updateReport deleteReport manageFolders updateDashboard publishDashboard listDataModelFiles readDataModelFile startDataModelEdit writeDataModelFile deleteDataModelFile getDataModelChanges getBranchDiff getDeploymentEnv getBranchState commitDataModelChanges commitToNewBranch mergeToDefaultBranch discardDataModelChanges switchUserBranch getPreAggregationStatus buildPreAggregation
dbt MCP Server
execute_sql text_to_sql get_dimension_values get_dimensions get_entities get_metrics_compiled_sql list_metrics list_saved_queries query_metrics get_all_macros get_all_models get_all_sources get_exposures get_lineage get_mart_models get_model_health get_model_performance get_node_details get_related_models search_public_mesh_models build clone compile docs get_lineage_dev get_node_details_dev list parse run show test cancel_job_run get_job_details get_job_run_artifacts get_job_run_details get_job_run_error list_job_run_artifacts list_jobs list_jobs_runs list_projects retry_job_run trigger_job_run generate_model_yaml generate_source generate_staging_model fusion.compile_sql fusion.get_column_lineage get_column_lineage get_product_doc_pages search_product_docs get_mcp_server_branch get_mcp_server_version
Read next
Sources
- docs.cube.dev/docs/integrations/mcp-server.md, read 10 Nov 2026
- github.com/dbt-labs/dbt-mcp/releases, read 10 Nov 2026
- docs.getdbt.com/docs/dbt-ai/about-mcp, read 10 Nov 2026
- raw.githubusercontent.com/dbt-labs/dbt-mcp/main/README.md, read 10 Nov 2026
- raw.githubusercontent.com/dbt-labs/dbt-mcp/main/.env.example, read 10 Nov 2026
Last reviewed 10 Nov 2026.