Notion
Index pages from Notion with an internal integration token. Connector type:
notion.
Setup
- Create a Notion internal integration and copy its token.
- Share the pages or the database to index with that integration.
- Create the connector with
credentials.tokenand, optionally,settings.database_id. Without a database ID the connector reads every page shared with the integration. - Start a sync with
POST /api/v1/connectors/{connector_id}/sync, or choose Sync now in the console.
What gets indexed
Each page becomes one document, written as Markdown:
- Headings (levels 1 to 3), paragraphs, bulleted and numbered lists, to-dos
- Quotes and callouts
- Code blocks, with their language
- Table rows
- Nested blocks (toggles, indented lists, columns), up to six levels deep
Long pages are read in full; the connector follows Notion's pagination for both the page list and each page's blocks. A child page is its own document. For a database, each entry's page content is indexed with its title; other property values are not. Comments, images and embedded files are not indexed. A page with no text is skipped and reported on the run. Archived and trashed pages are out of scope.
Sync behaviour
- Incremental. A sync reads pages edited since the last applied sync started.
- Deletions. A page that is archived, trashed or no longer shared with the integration is deleted from the index. A sync started with
limitdoes not remove anything. - Schedule. A stored cron schedule is executed. The catalogue suggests every six hours (
0 */6 * * *). - Refusals. If Notion rejects the token or the page listing, the run fails with
source_refused.
Access
Notion sharing is not mirrored. Everything the integration can read is available to everyone who can use the project, so invite the integration only to pages meant for that audience. There is no hosted OAuth consent flow.