Cube Cloud MCP server

What the model can do

Cube Cloud semantic layer: deployments, the data model's measures and dimensions, queries against it, workbooks and dashboards, reports, AI scheduled tasks, dashboard notification schedules and build status; when you ask, can create, change, run and delete scheduled tasks, schedule dashboard notifications and add or remove their recipients, create, copy, publish and delete workbooks, refresh and delete reports, and run report deliveries.

Reads

• cube_deploymentsCube Cloud deployments (each a semantic layer over a data source): id, name, address and setup step.
• cube_modelThe deployment's semantic model: cubes and views with their measures, dimensions and segments (names, titles, types).
• cube_queryRuns a query against the deployment's semantic layer (Cube's JSON query) and returns the rows: measures aggregated by dimensions…
• cube_workbooksWorkbooks (each holding reports and an optional published dashboard) in a deployment: id, name, folder, whether published, dates.
• cube_reportsReports (saved queries) in a deployment, most recently updated first: id, public id, name, title, description, workbook, folder, dates.
• cube_scheduled_tasksAI scheduled tasks in a deployment: the prompt the agent runs, its cron and timezone (or manual), whether enabled, next run…
• cube_notificationsScheduled dashboard notifications in a deployment (a dashboard rendered and emailed on a schedule): id, dashboard, schedule in words and cron…
• cube_build_statusThe latest build of a deployment's data model (production, or a branch): status, deployment status, any error text and when it changed.

Acts, when you ask

These change something in Cube Cloud. The model is told to use them only when you ask for exactly that, and each one is granted by name: per team on a shared connection, by the owner on a personal one.

• cube_create_scheduled_taskCreates an AI scheduled task in Cube: the agent runs the prompt as the key's user on the cron schedule (at most hourly…
• cube_update_scheduled_taskChanges a Cube scheduled task: its name, prompt, description, timezone, schedule (cron, or manual: true to run only on demand), or pauses…
• cube_delete_scheduled_taskPermanently deletes a Cube scheduled task and its schedule; the chat threads of past runs are kept.
• cube_run_scheduled_taskRuns a Cube scheduled task once now, even if paused or manual: the agent queries the warehouse (compute cost) and, if the task notifies…
• cube_create_notificationSchedules a Cube dashboard notification: the published dashboard is rendered on the schedule (hourly, daily, weekly, monthly…
• cube_add_notification_recipientsSubscribes Cube console users, by user id or email, to a dashboard notification: from then on each scheduled render is emailed to them.
• cube_remove_notification_recipientsUnsubscribes Cube console users, by user id, from a dashboard notification, so they stop receiving it.
• cube_delete_notificationPermanently deletes a Cube dashboard notification and all its recipients; nobody receives it again.
• cube_create_workbookCreates an empty Cube workbook, owned by the key's user, optionally in a folder.
• cube_duplicate_workbookCopies a Cube workbook with its reports and published dashboard, as "Copy of <name>", owned by the key's user; sharing and embedding are not copied.
• cube_delete_workbookPermanently deletes a Cube workbook, its reports and its dashboard; anyone using that dashboard loses it.
• cube_publish_dashboardPublishes a Cube workbook's saved dashboard draft, so everyone with access to the dashboard sees the new version in place of the old.
• cube_refresh_reportRefreshes a Cube report, re-running its query against the warehouse (compute cost) and updating its stored result.
• cube_delete_reportPermanently deletes a Cube report.
• cube_run_report_deliveryRuns a Cube report delivery now: the report's current results are written into its Google Sheets or Excel destination (overwriting…

Connect it

First connect Cube Cloud on Dashboard → Connections, then add Memtro to your client once. The Cube Cloud tools appear alongside the tools of everything else you have connected; nothing is configured per connector.

claude plugin marketplace add memtro-org/claude-memtro
claude plugin install memtro@memtro --config api_key=sk-memtro-...

{
  "mcpServers": {
    "memtro": {
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

{
  "mcp": {
    "memtro": {
      "type": "remote",
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

Who may use it

The key identifies the person, and the server offers that person the tools they may use everywhere else in Memtro. A shared Cube Cloud connection is used by the teams an administrator granted it to, tool by tool, on the Permissions page; a personal connection is used by its owner alone, and its tools that act stay off until the owner allows them. See Teams and permissions and Connectors.

Ask it like this

• › What was our order count by status last month in Cube?
• › Which measures does the orders cube have?
• › Did the latest build of the production model succeed?
• › Which AI scheduled tasks run every morning, and who do they notify?
• › Refresh the weekly revenue report.

Questions

Is this an official Cube Cloud MCP server?

No. It is Memtro's MCP server with Memtro's Cube Cloud connector behind it: the connector reads Cube Cloud through its API with the credentials you connect on Dashboard → Connections, and Cube Cloud publishes nothing here. The same endpoint carries every other system you connect and Memtro's memory tools.

What does it cost?

Nothing on top of Memtro. One price per organisation by its number of users, from $50 a month for up to 10 users after a 7-day free trial, with every connector, the MCP server, jobs and memory included. You bring your own model keys.

Can it change anything in Cube Cloud?

Only when you ask, and only what it has been granted: create an AI scheduled task; change, pause or resume a scheduled task; delete a scheduled task; run a scheduled task now; schedule a dashboard notification; add recipients to a dashboard notification; remove recipients from a dashboard notification; delete a dashboard notification; create a workbook; copy a workbook; delete a workbook; publish a workbook's dashboard draft; refresh a report; delete a report; run a report delivery to its spreadsheet now. Every other tool reads, searches or lists. An administrator grants each acting tool by name per team, and a personal connection acts only after its owner allows it.

Which clients work?

Any MCP client that speaks Streamable HTTP: Claude Code, Cursor, Claude Desktop, OpenCode, Codex, Zed, Windsurf, Cline and the Anthropic and OpenAI SDKs. There is no local process to run; the server is remote and authenticated with your key.

Other MCP servers · Databases and warehouses

• Postgres MCP server
• MySQL and MariaDB MCP server
• SQL Server MCP server
• BigQuery MCP server
• Snowflake MCP server
• AWS Athena MCP server