Adds a comment to a Notion page, to a block, or as a reply in an existing discussion. Takes exactly one of page_id, block_id and discussion_id, and the text as plain text. A new comment on a selection of text is not possible through Notion's API; a reply to an existing one is, by its discussion_id. Returns the comment's id, discussion_id, created_time, created_by, text and parent.

Creates a Notion database under a page, with one data source and that data source's property schema. The schema is a map of property name to type, for example {"Task": "title", "Due": "date", "Stage": {"type": "status", "options": ["Todo", "Done"]}}. A map with no title property gets one named Name. Returns the database's id, title, url, parent, is_inline and data_sources (their ids are what notion_create_page and notion_query_data_source take).

Creates one Notion page: under a page, or as a row of a data source. Takes the parent, a title, property values as plain values (for a row), content as markdown, and an optional template. For a row, the data source schema is read once and each value is encoded against it. Returns the new page's id, url, title, parent and properties. A failed create is not sent again; a create that got no answer may still have happened, and notion_search by title shows whether it did. Content and a template are not taken together.

Adds a view to a Notion database: a table, board, list, calendar, timeline, gallery, form, chart, map or dashboard over one of its data sources, with an optional filter, sorts and layout configuration in Notion's JSON. Returns the view's id, name, type, database_id, data_source_id, url, filter and sorts. The view id is what notion_query_data_source takes as view_id.

Deletes one Notion comment. Notion's API deletes only comments created through the same connection, and answers not found for any other. Returns the comment id and deleted true.

Copies one Notion page under a parent page or into a data source, by reading the source page and creating a new one; Notion's API has no duplicate. The copy carries the title, the content as markdown, the icon and cover when they are an emoji or a link, and, in a data source, the property values the target schema accepts. Comments, page history, page-level permissions, files stored in Notion, and blocks with no markdown form are not carried by a duplicate; content is copied as Notion's markdown rendering, so some block formatting can differ. Returns the new page's id and url, properties_not_copied with the reason for each, unknown_block_ids (blocks that could not be copied) and content_truncated.

Reads the open comments on one Notion page or block. Each comment is reduced to id, discussion_id (comments in one thread share it), created_time, created_by (a name, or an id when Notion gives no name), text and parent. Resolved comments are not returned by Notion's API. The result is {results, has_more, next_cursor}, with the same shape when there are no results. has_more true means more results exist than this page holds, and next_cursor is the cursor argument that returns them.

Reads one Notion database and the schema of each of its data sources. In Notion's current API a database is a container of one or more data sources, and rows and properties live on the data source; most databases have one. Returns the database's id, title, url, parent and in_trash, and data_sources: for each, its id, name and properties as a list of {name, type}, with options (the option names) for select, multi_select and status properties. The data source id is the one notion_query_data_source takes.

Reads one Notion page: its title, url, parent, timestamps and properties, and its content as markdown. Properties come back as plain values keyed by property name, for example {"Status": "Done", "Assignee": ["Test User"], "Due": "2026-10-20"}. A database row is a page, so a row's id returns that row. The markdown is Notion's own rendering; truncated is true when Notion cut it short (very long pages), and unknown_block_ids lists blocks with no markdown form, such as embeds and link previews. A page that is not shared with the DataToRAG connection reads as not found.

Lists the people and bots in the Notion workspace, or reads one of them. Each user is reduced to id, type (person or bot), name, and email for a person when Notion includes it; a bot also carries workspace_name. With user_id set to "me", returns the DataToRAG connection's own bot user, whose workspace_name names the workspace this connection is in. The result is {results, has_more, next_cursor}, with the same shape when there are no results. has_more true means more results exist than this page holds, and next_cursor is the cursor argument that returns them. A single user comes back in the same shape, as a one-item results list.

Moves one Notion page to a new parent: another page, or a data source (where it becomes a row). Pages only; Notion's API does not move databases or other blocks. Returns the page's id, url and new parent.

Reads the rows of one Notion data source (a table inside a database), with an optional filter, sort order, property allowlist, page size and cursor. Each row is {id, url, created_time, last_edited_time, properties}, with properties as plain values keyed by property name. Filter and sorts are passed to Notion as written, in Notion's query JSON; notion_get_database lists the property names and types they refer to. The result is {results, has_more, next_cursor}, with the same shape when there are no results. has_more true means more results exist than this page holds, and next_cursor is the cursor argument that returns them. With view_id in place of data_source_id, the rows are a saved view's, in its own filter and sort order, and each row is its id alone, which is what Notion's view query returns; notion_get_page reads a row's properties. A view query's results stay cached for 15 minutes; total_count is the number of rows it matched, and incomplete true means Notion's 10,000-row cache cut them short.

Searches the Notion pages and data sources shared with the DataToRAG connection, by title. Notion's search matches titles, not page content, and Notion describes its index as not immediate and not guaranteed complete, so a page created moments ago can be missing. Takes an optional query, an optional object type, a page size and a cursor. Each result is reduced to id, object (page or data_source), title, url, parent ({type, id}) and last_edited_time. The result is {results, has_more, next_cursor}, with the same shape when there are no results. has_more true means more results exist than this page holds, and next_cursor is the cursor argument that returns them. A page or database that exists but is not shared with the connection does not appear.

Replaces the text of one Notion comment. Notion's API edits only comments created through the same connection, and answers not found for any other. Returns the comment.

Changes one Notion data source (the table inside a database): its title, its property schema, or whether it is in the trash. The schema change is a map keyed by a property's current name: null removes the property for every row, {"name": "New"} renames it, and a type (alone or with name) changes its type and settings, in the forms notion_create_database takes. A select's options list replaces the whole list. Changing a type can leave existing values unreadable in the new type. Returns the data source's id, title, in_trash and property schema.

Changes one Notion page's title, property values, icon or cover, applies a template, or moves it to or from the trash. Property values are plain values checked against the page's data source schema (one read of the page and one of its schema). in_trash true is how a page is deleted through Notion's API, and in_trash false restores it. Omitted properties keep their values; a list value replaces the whole list. Returns the page's id, url, title, parent, in_trash and properties.

Edits one Notion page's content as markdown, in one of three modes. edits replaces exact passages, each {old_str, new_str, replace_all_matches}; an edit whose old_str matches nothing is listed in warnings while the others apply. replace_with replaces the whole content. insert adds markdown at the start, at the end (the default), or after a selection. Child pages and databases are kept unless allow_deleting_content is true, and Notion refuses an edit that would remove them. The content of a synced page is not edited through Notion's API. Returns the page id and Notion's warnings.

Changes one Notion database view's name, filter, sorts or layout configuration. null clears the filter or the sorts; omitted fields keep their values. A view's type is not changed through Notion's API. Returns the view in the shape notion_create_view returns.