Integration

Cube Cloud integration

Cube Cloud is a hosted semantic layer and BI platform. With its Platform API an agent can list deployments and their build status, manage workspace folders, reports and workbooks (dashboards as code: upsert by publicId or slug, edit definitions, publish, promote to another deployment), export dashboards as PNG or PDF, manage data model branches (dev mode, delete or rename files, commit, merge, pull, validate), connect a GitHub repository, and run dbt syncs, Databricks Metric View and Snowflake Semantic View publication and Apache Ossie conversions. Not exposed: API keys, deployment and environment tokens, environment variables (database credentials), OAuth and OIDC configuration, users, roles changes, user attributes, SCIM, embedding and embed tenants, chat history, runtime logs and pods, tenant settings, scheduled notifications, scheduled tasks, report deliveries, spreadsheet placements, data model file listing and writing file content (model code runs inside the deployment with its database credentials), deployment creation, update and deletion, and the Core Data and AI APIs (separate hosts and tokens). Creator emails, names and pictures are not returned. Authentication is a Cube Cloud API key sent as a Bearer token to your tenant host; it acts with its owner's permissions.

Connect Cube Cloud How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Cube Cloud.Built from the 92 Cube Cloud tools below, on the connection you add once.

What you can do with Cube Cloud

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

File, 6 operations

delete_data_model_filesDelete data model filesDELETEDeletes data model source files from a branch (by path). The change is pending until commit_changes.
get_dashboard_draft_fileGet dashboard draft fileGETReturns a workbook's dashboard draft as canonical JSON text (content) plus its hash. Edit it with replace_dashboard_draft_text, passing the hash as expectedHash.
get_ossie_conversion_filesGet ossie conversion filesGETReturns the files a completed Apache Ossie conversion produced (ossie.yaml for an export, Cube model files for an import), each with path and content. Pages with first and after.
get_report_fileGet report fileGETReturns a report definition as canonical JSON text (content) plus its hash. Edit it with replace_report_file_text, passing the hash as expectedHash.
replace_dashboard_draft_textReplace dashboard draft textPATCHEdits a dashboard draft by exact text replacement (oldText to newText, in order); all edits apply or none. Read get_dashboard_draft_file first; with expectedHash a changed draft gives 409.
replace_report_file_textReplace report file textPATCHEdits a report definition by exact text replacement (oldText to newText, in order); all edits apply or none. Read get_report_file first; with expectedHash a changed file gives 409.

Report, 5 operations

create_reportCreate reportPOSTCreates a report: a saved query (sqlQuery or jsonQuery) with its pivot and visualization. Returns the report with its numeric id, which dashboard widgets reference.
delete_reportDelete reportDELETEDeletes a report by its numeric id.
get_reportGet reportGETReturns a report's full definition: name, title, sqlQuery or jsonQuery, pivotItems, visualization meta, folder, workbook, publicId and the roles with access.
list_reportsList reportsGETLists the reports (saved queries with their visualization) of a deployment, filtered by workbook, folder or a name search. Pages with first and after.
update_reportUpdate reportPUTUpdates a report with the fields sent: name, title, query, pivot, visualization meta, folder or workbook. Can also assign a publicId to a report that has none (write-once).

Result, 5 operations

get_databricks_preview_resultGet databricks preview resultGETReturns a completed Databricks Metric View preview: per view, the generated Metric View YAML, counts, issues and the measures it could not represent.
get_dbt_sync_resultGet dbt sync resultGETReturns the result of a completed dbt sync: cube count, duration, manifest statistics and the generated file paths.
get_ossie_conversion_resultGet ossie conversion resultGETReturns the result of an Apache Ossie conversion: outcome, element counts and the issues the converter raised (ERROR, WARNING, INFO).
get_snowflake_pull_resultGet snowflake pull resultGETReturns the result of a completed Snowflake Semantic View pull: view count, duration and the generated file paths.
get_snowflake_push_resultGet snowflake push resultGETReturns the result of a completed Snowflake Semantic View push.

Workbook, 5 operations

create_workbookCreate workbookPOSTCreates a workbook (a dashboard container) in a deployment, optionally in a folder and with metadata such as meta.dashboardDraft.
delete_workbookDelete workbookDELETEDeletes a workbook (the container that holds a dashboard) by its numeric id.
get_workbookGet workbookGETReturns a workbook with its dashboard draft and published dashboard (widgets, layout, theme), folder path, slug and promotion provenance.
list_workbooksList workbooksGETLists the workbooks of a deployment (containers holding a dashboard), optionally in one folder. Pages with first and after.
update_workbookUpdate workbookPUTUpdates a workbook's name, folder, slug or metadata. Only the fields sent change. slugTakeover moves a slug already used by another workbook.

Branch, 4 operations

create_branchCreate branchPOSTCreates a data model branch from another branch (from), optionally switching dev mode to it (activate).
delete_branchDelete branchDELETEDeletes a data model branch; child branches move to its parent. The deploy branch cannot be deleted (400). removeOnUpstream true also deletes it on the remote git repository (default false).
list_branchesList branchesGETLists the data model branches of a deployment with their parent branch, last commit hash and staging environment flag. Pages with first and after.
list_github_branchesList github branchesGETLists the branches of a GitHub repository through a GitHub App installation, flagging the default branch.

Dbt sync, 4 operations

cancel_dbt_syncCancel dbt syncDELETECancels a running dbt sync.
get_dbt_sync_statusGet dbt sync statusGETReturns a dbt sync's status and progress. COMPLETED and FAILED are the only terminal values; then read get_dbt_sync_result, or get_dbt_sync_generated_files when nothing was committed.
list_dbt_syncsList dbt syncsGETLists the dbt syncs of a deployment, optionally filtered by status or trigger. Pages with first and after.
start_dbt_syncStart dbt syncPOSTStarts a dbt sync that generates Cube data model files from a dbt project (git clone or an uploaded manifest). Returns syncJobId; poll get_dbt_sync_status until COMPLETED or FAILED.

Folder, 4 operations

create_folderCreate folderPOSTCreates a workspace folder in a deployment, at the root or inside parentId.
delete_folderDelete folderDELETEDeletes an empty workspace folder. A folder that still has sub-folders answers 400 (delete or move them first). Workbooks and reports directly inside it are not deleted: they move back to the workspace root.
list_foldersList foldersGETLists the workspace folders of a deployment, or only the children of parentId. Pages with first and after.
update_folderUpdate folderPUTRenames or reorders a workspace folder. Only the fields sent are changed: leave a field out to keep its current value.

Sync, 4 operations

cancel_databricks_publication_runCancel databricks publication runDELETERequests cancellation of a running Databricks Metric View publication run.
get_databricks_publication_runGet databricks publication runGETReturns a Databricks Metric View publication run: status, progress and the outcome of each view, including measures withheld.
list_databricks_publication_runsList databricks publication runsGETLists the Databricks Metric View publication runs of a data source, filtered by status. Pages with first and after.
start_databricks_publication_runStart databricks publication runPOSTStarts publishing Cube views as Databricks Metric Views. Real effect: creates or updates views in Databricks. Returns runId; poll get_databricks_publication_run.

Databricks metric view preview, 3 operations

cancel_databricks_previewCancel databricks previewDELETECancels a running Databricks Metric View preview.
get_databricks_preview_statusGet databricks preview statusGETReturns a Databricks Metric View preview's status and progress; resultAvailable says when get_databricks_preview_result can be read.
start_databricks_previewStart databricks previewPOSTStarts a dry run that converts Cube views to Databricks Metric View YAML without publishing anything. Returns previewId; poll get_databricks_preview_status, then get_databricks_preview_result.

Ossie conversion, 3 operations

cancel_ossie_conversionCancel ossie conversionDELETECancels a running Apache Ossie conversion.
get_ossie_conversion_statusGet ossie conversion statusGETReturns an Apache Ossie conversion's status: QUEUED, RUNNING, COMPLETED (result available), FAILED or CANCELLED.
start_ossie_conversionStart ossie conversionPOSTStarts converting between a Cube data model and an Apache Ossie model (direction export or import). Returns conversionId; poll get_ossie_conversion_status until COMPLETED, FAILED or CANCELLED.

Promotion, 3 operations

get_workbook_promotionGet workbook promotionGETReturns where a workbook was promoted from and to: its origin, its parent and its first and last promotion.
list_workbook_promotionsList workbook promotionsGETLists the promotion history of a workbook: source, target, outcome and when. Pages with first and after.
promote_workbookPromote workbookPOSTPromotes a workbook to another deployment and publishes it there. Real effect: the target dashboard is replaced. Run preview_workbook_promotion first for the hashes.

Dashboard export, 2 operations

get_dashboard_export_statusGet dashboard export statusGETReturns the status of a dashboard export: pending or processing (keep polling), completed (download it with download_dashboard_export) or failed (see error).
start_dashboard_exportStart dashboard exportPOSTStarts rendering a dashboard as PNG or PDF and returns a jobId. Poll get_dashboard_export_status until completed, then call download_dashboard_export before expiresAt.

Databricks metric view integration, 2 operations

get_databricks_publication_settingsGet databricks publication settingsGETReturns the Databricks Metric View publication settings of one data source: target catalog and schema, view selection, naming, and configurationVersion (needed to save or start a run).
save_databricks_publication_settingsSave databricks publication settingsPUTCreates or replaces the Databricks Metric View publication settings of a data source. Pass expectedConfigurationVersion 0 to create, else the configurationVersion from get_databricks_publication_settings.

Deployment, 2 operations

get_deploymentGet deploymentGETReturns one deployment: its id, name, setup step (creationStep) and the URL of its data API.
list_deploymentsList deploymentsGETLists the Cube Cloud deployments the API key can reach, with id, name, API URL and setup step. Use an id as deploymentId in every other tool. Pages with first and after (pageInfo.endCursor).

Dev mode, 2 operations

start_dev_modeStart dev modePOSTEnters dev mode on a deployment, switching to branchName, so data model edits go to that branch.
stop_dev_modeStop dev modeDELETEExits dev mode for the caller on a deployment.

Role, 2 operations

get_roleGet roleGETReturns one custom role with its policies (resource type, resources and allowed actions).
list_rolesList rolesGETLists the account's custom roles with their policies (resource type, resources and allowed actions). Pages with first and after.

Snowflake semantic view pull, 2 operations

get_snowflake_pull_statusGet snowflake pull statusGETReturns the status and progress of a Snowflake Semantic View pull. COMPLETED and FAILED are terminal; then read get_snowflake_pull_result.
start_snowflake_pullStart snowflake pullPOSTStarts importing Snowflake Semantic Views from a schema as Cube data model files written to a branch. Returns syncJobId; poll get_snowflake_pull_status until COMPLETED or FAILED.

Snowflake semantic view sync, 2 operations

get_snowflake_push_statusGet snowflake push statusGETReturns the status and progress of a Snowflake Semantic View push. COMPLETED and FAILED are terminal; then read get_snowflake_push_result.
start_snowflake_pushStart snowflake pushPOSTStarts publishing Cube views to Snowflake as Semantic Views. Real effect: creates or replaces views in the Snowflake schema. Returns syncJobId; poll get_snowflake_push_status.

Agent default, 1 operation

get_agent_defaultsGet agent defaultsGETReturns a deployment's AI agent defaults: whether the automatic agent is on and the default agent id.

Agent suggestion, 1 operation

list_agent_suggestionsList agent suggestionsGETLists the suggested questions shown for a deployment's AI agent, as plain strings. Pages with first and after.

Ancestor, 1 operation

list_folder_ancestorsList folder ancestorsGETLists the parent folders of a folder up to the root, to build its path.

App config, 1 operation

get_app_configGet app configGETReturns the account's app configuration: the theme (palette, fonts, logos), feature switches for AI and dashboards, embedding options and time zone settings.

Build status, 1 operation

get_build_statusGet build statusGETReturns the latest build or compile status of a branch (the production build, or dev mode), with its status, error text and file hashes. Leave branchName out for the deploy branch.

Bulk delete, 1 operation

bulk_delete_workspace_itemsBulk delete workspace itemsPOSTDeletes up to 100 folders, workbooks or reports in one call. Partial success: deleted lists what was removed, failed lists the rest with the error each one got.

Bulk move, 1 operation

bulk_move_workspace_itemsBulk move workspace itemsPOSTMoves several folders, workbooks or reports into one folder (or the root when folderId is left out).

By public id, 1 operation

upsert_report_by_public_idUpsert report by public idPUTCreates or updates a report keyed by a publicId you choose (12 characters, [0-9A-Za-z], unique in the account). Re-applying the same definition is a no-op; 409 on conflict.

By slug, 1 operation

upsert_workbook_by_slugUpsert workbook by slugPUTCreates or updates a workbook keyed by a slug you choose, carrying its dashboard layout in meta.dashboardDraft. Only fields sent change; meta is merged. Publish it with publish_dashboard.

Commit, 1 operation

commit_changesCommit changesPOSTCommits and pushes the pending data model changes of the branch active in dev mode. Real effect: writes to the repository.

Connect, 1 operation

connect_github_repositoryConnect github repositoryPOSTConnects a deployment to a GitHub repository (existing, or created when isNewRepo is true) and triggers a build. Real effect: the deployment then builds from that repository.

Dashboard, 1 operation

update_workbook_dashboardUpdate workbook dashboardPUTReplaces a workbook's dashboard draft (dashboardDraft: widgets, layout, theme). Viewers see it only after publish_dashboard.

Download, 1 operation

download_dashboard_exportDownload dashboard exportGETDownloads the rendered PNG or PDF of a completed dashboard export as a file.

Duplicate, 1 operation

duplicate_workbookDuplicate workbookPOSTCreates a copy of a workbook with its dashboard and returns the new workbook.

Generated file, 1 operation

get_dbt_sync_generated_filesGet dbt sync generated filesGETReturns the cube files a dbt sync generated without committing them (output files), each with path and YAML content. Pages with first and after.

Installation, 1 operation

list_github_installationsList github installationsGETLists the GitHub App installations (organizations or accounts) linked to the caller, with the installationId used by list_github_repositories and connect_github_repository.

Merge, 1 operation

merge_into_parent_branchMerge into parent branchPOSTMerges a branch into its parent branch (git-flow merge), optionally squashing commits and deleting the branch. Real effect: writes to the repository.

Merge to default, 1 operation

merge_to_default_branchMerge to default branchPOSTMerges a branch straight into the deploy (default) branch with a commit message, optionally removing the branch. Real effect: changes what production builds.

Move, 1 operation

move_workspace_itemMove workspace itemPOSTMoves one folder, workbook or report into a folder (or the root when folderId is left out).

Promotion link, 1 operation

list_workbook_promotion_linksList workbook promotion linksGETLists the destinations (deployment and workbook) a workbook has been promoted to. Pages with first and after.

Promotion preview, 1 operation

preview_workbook_promotionPreview workbook promotionPOSTPreviews promoting a workbook to another deployment without changing anything: returns sourceHash and destinationHash to pass to promote_workbook, the report count and whether the target diverged.

Publish, 1 operation

publish_dashboardPublish dashboardPOSTPublishes a workbook's dashboard draft so viewers see it. Real effect: the published dashboard changes for everyone. A dashboard that requires review answers 400.

Pull, 1 operation

pull_branchPull branchPOSTPulls the latest commits of a branch from its remote repository and rebuilds when something changed (buildTriggered).

Region, 1 operation

list_regionsList regionsGETLists the regions a Cube Cloud deployment can run in, with provider and whether each is public, dedicated or hybrid.

Rename, 1 operation

rename_data_model_filesRename data model filesPOSTRenames or moves data model source files on a branch (path to newPath). The change is pending until commit_changes.

Repository, 1 operation

list_github_repositoriesList github repositoriesGETLists the repositories a GitHub App installation can access. Pages with first and after.

Settings, 1 operation

get_deployment_settingsGet deployment settingsGETReturns a deployment's settings: cloud provider and region, deploy branch and project root, Cube version and release channel, custom domain, AI agent defaults and launchpad tabs.

Shared workspace, 1 operation

list_shared_workspace_itemsList shared workspace itemsGETLists the folders, workbooks and reports of a deployment's shared workspace, optionally in one folder, filtered by type or name. Pages with first and after.

Staging environment, 1 operation

set_branch_staging_environmentSet branch staging environmentPUTTurns a branch's staging environment on or off. Name the branch by branchName or branchId.

Validate, 1 operation

validate_data_modelValidate data modelGETCompiles a deployment's data model (a branch, or dev mode) and returns whether it is valid, the cube count and each compilation error with its file.

Version, 1 operation

list_deployment_versionsList deployment versionsGETLists the Cube versions available to a deployment, flagging the current one and the latest of each release channel.

Workspace, 1 operation

list_workspace_itemsList workspace itemsGETLists the folders, workbooks and reports of the caller's workspace in a deployment, optionally in one folder, filtered by type or name. Pages with first and after.

Use Cube Cloud from Claude Code, Cursor or Codex

Connect Cube Cloud in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Cube Cloud operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Cube Cloud

Create your workspace
Sign up free. Nothing to install.
Connect Cube Cloud once
Add your Cube Cloud connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.