Skip to main content

Data Sources

Data sources are governed remote connections to your data — the databases and files that power the rest of MINEO. Once connected, Threads and Assistants query them in natural language, notebooks read from them, and data observability scanners run checks against them. You configure each connection once, with access controlled per user and group.

Currently supported data sources are:

  • PostgreSQL (Including AWS Redshift)
  • ClickHouse
  • Oracle
  • MSSQL
  • MySQL
  • CSV Files
  • SQLite Files
  • DuckDB Files
note

We are progressively including more datasources, so visit this page often, or read our blog to be informed about changes. If you want to work with a datasource that is not supported in this list, please contact us.

Organization-level data sources​

Organization administrators can share a data source with the whole organization, making it available in every project.

  • Share: In the project's Data Sources list, open the ... menu of the data source and select Visibility → Organization. The data source then appears in every project's Data Sources list and data source selectors, and in the organization's Workspace → Data Sources tab.
  • Stop sharing: Select Visibility → Project from the same menu, or open the ... menu on the data source in the organization's Workspace → Data Sources tab and select Project. The data source returns to its origin project and disappears from the other projects.
  • Badge: Shared data sources display an Organization badge in lists and selectors.

The data source always keeps its origin project. Users in other projects can query it from Threads and Assistants, but cannot edit, clone, or delete it; those actions stay in the origin project. Cloning it from the origin project produces a project-private copy.

warning

Sharing a data source grants query access to every organization member and viewer, not only to the members of the origin project.

  • For server databases (PostgreSQL, ClickHouse, Oracle, MSSQL, MySQL), use a read-only database user when sharing.
  • For file-based data sources (CSV, SQLite, DuckDB), the file is read through the origin project's storage regardless of the file's own permissions — sharing the data source shares the file's content.

When sharing stops, the data source disappears from the other projects' lists and selectors. Assistants in those projects that already had it connected keep querying it; disconnect it from their Knowledge tab if that is not intended. The organization's Workspace → Data Sources tab asks for confirmation before stopping sharing.

See also​