Menu

Supported Operations

Supported operations

SqlBak MCP exposes your SqlBak account to an MCP client (e.g. Cursor, Claude Desktop) via a set of tools. The tools are not limited to read-only inspection. In practice, SqlBak MCP supports read-only, create, update, validation, and guided restore flows.

Servers (agents)

  • get_servers — List all servers (agents): name, OS, status (online/offline/paused), version, timezone, down-alert settings. Optional filter by server name.
  • get_agent_install_instructions — Installation steps for the SqlBak agent: Windows (MSI), Linux (.deb, .rpm), Docker. Includes download links, versions, and a secret key for agent registration. Optional filters: platform, secret key name.

Jobs

  • get_jobs — List all backup, restore, and maintenance jobs with summary info: server, sources, storages, notifications, schedule, last run status, daily stats. Optional filter by server.
  • get_job_details — Full configuration and paginated run history for a single job by jobId.
  • get_job_run_details — Full details for a single run by jobRunId: configuration, status, paginated logs with optional level filter (debug, warning, error, information).
  • get_default_job_settings — Default settings template for creating a job. Useful to inspect defaults for a job type, agent, and optionally a specific source.
  • create_job — Create a new backup or maintenance job: name, agent, sources, storages, schedule, notifications, and options (compression, encryption, handlers, and more). Parameters follow the schemas from reference data and default job settings.
  • edit_job — Update an existing backup or maintenance job. This is the main tool for scenarios such as changing schedule, changing retention, and adding failure notifications.

Important notes:

  • a backup job allows only one source
  • a backup job requires at least one storage
  • a maintenance job requires at least one non-empty maintenanceHandlers.text
  • encryption depends on the current service plan

Sources and storages

  • get_sources — List data sources available for backup. Optional filters: agentId, sourceTypeId (mssql, mysql, pgsql), sourceName.
  • get_source_metainfo — Detailed metadata for one source by sourceId: list of databases with properties (system vs user, backup enabled/disabled), and for SQL Server, database file details. Use when you need exact database names or file info before creating a job.
  • get_storages — List backup destinations (e.g. Amazon S3, Azure Blob, SFTP, folder). Optional filters: agentId, storageTypeId, storageName.
  • get_storage_details — Detailed configuration for a single storage by id. Use when you need the exact storage settings before testing or updating it.
  • create_or_update_storage — Create a new storage or update an existing one.
  • authorize_storage_oauth — Start OAuth authorization for storage types such as Google Drive, Dropbox, Box, OneDrive, or Yandex Disk. Returns a direct authorization URL and a reusable requestId.
  • test_storage — Validate connectivity and basic access to a storage. This can be used with a new inline configuration or after reading an existing storage configuration.

Note: get_source_metainfo is a slower API call. Use it only after you have the source list from get_sources and need detailed database or file information.

Note: For an existing storage, the recommended flow is get_storages → get_storage_details → test_storage.

Reference data and defaults

  • get_reference_data — Reference metadata: storage/source/job/backup types, compression options, archive options, event handler types, service-plan capabilities, and validation rules for fields used in other tools.
  • get_default_job_settings — Default settings template for creating a job: defaults for job type, agent, and optionally a specific source. Useful to pre-fill or validate parameters for create_job.

Notifications and billing

  • get_notification_channels — List notification channels (e.g. Email, Slack) available for job alerts.
  • get_notification_channel_details — Detailed type-specific settings for one notification channel by id.
  • create_or_update_notification_channel — Create a new notification channel or update an existing one.
  • test_notification_channel — Send a one-off test notification without creating or updating a persistent notification channel.
  • get_pricing — Subscription plans and pricing for SqlBak.

Restore

  • run_restore_job — Start a restore from an existing successful backup run using a guided multi-step flow. The flow requires explicit confirmation at every step and should not be auto-started by the AI without user confirmation.

Important note: run_restore_job is confirmation-driven and destructive by nature. The AI should present the available options, ask for confirmation, and only then continue to the next restore step.