Beta: Live Mode (Data fetching) is in beta and may change.
- Data Crawling and Indexing: A workspace-level integration that builds a shared index of content explicitly shared with Glean.
- Live Mode (MCP-based): A real-time data fetcher that retrieves fresh, permission-aware content at query time via per-user OAuth.
Supported Features
The connector enables central ingestion and search for shared organizational content.- Comprehensive Retrieval: Ingests Notion pages and databases, including full recursive crawling of nested “block” content such as toggles, tables, and task trackers.
- Metadata Support: Indexes page/database titles, owner/author information, and last-modified timestamps.
- Flexible Data Modes:
- Data Crawling: Uses an internal integration token to index content on a regular schedule.
- Live Mode: Uses the Model Context Protocol (MCP) to provide real-time retrieval, ensuring Assistant answers include private or recently updated content without waiting for a crawl.
Supported Objects
- Pages
- Databases
- Blocks (toggle blocks and sub-blocks) within pages and databases are indexed recursively to a configurable depth, allowing search over deep Notion structures.
Supported Blocks
- paragraph
- heading_1
- heading_2
- heading_3
- bulleted_list_item
- numbered_list_item
- to_do
- toggle
- child_page
- child_database
- image
- embed
- video
- file
- bookmark
- text
- unsupported
- code
- callout
- quote
- divider
- equation
- table_of_contents
- column_list
- breadcrumb
- synced_block
- link_to_page
- link_preview
- template
- table_row
Supported Page Properties
Notion page properties are automatically indexed and searchable in Glean, allowing you to search and filter documents based on custom page properties. Supported Properties: title, rich text, people, email, status, number, date, select, checkbox, multi-select How to Search:- Direct search: Type any property value in the search box to find pages containing that value
- Filtered search: Type the property name, colon, and value (e.g.,
assignee:"Steve Smith"orstatus:Done) to filter by specific properties
startdate:2024-01-01, “due_date” becomes duedate:2024-12-31)
Crawling Strategy
Notion uses scheduled full crawls (no webhooks today) and a separate identity crawl.| Crawl type | Scope | Frequency | Notes |
|---|---|---|---|
| Full content crawl | Pages, databases, blocks shared with the integration | Every 6 hours | No webhooks; used to pick up new/updated content and structural changes. |
| Identity crawl | Users/workspaces relevant to Notion | Every hour | Keeps user/identity maps fresh for permissions modeling outside Notion. |
| Deletion safety | Previously indexed docs | 10‑day delay | Docs are only hard-deleted from the index if they remain missing for >10 days. |
How it Works
1
Notion workspaces are installed
You install the Glean notion connector individually for each Notion workspace. You share each page with the integration that you wish to index. Child pages of shared pages will automatically be shared with the integration.
2
Glean crawls the Notion workspace
Glean will crawl the Notion workspace and index all pages and their child pages that you’ve shared with the integration.
3
Crawled content is made searchable
Glean indexes the crawled content and makes it searchable for all Glean users who can access the Notion connector. Because Notion’s crawl APIs do not expose document‑level permission lists in a way Glean can project into its index, this indexed corpus is not permission‑aware—any content shared with the integration is visible in Glean to all such users.
4
(Optional) Live Mode fetches permission‑aware content at query time
When Live Mode is enabled and a user has connected their Notion account via OAuth, Glean Assistant can call Notion via MCP at query time to retrieve fresh, permission‑aware content on behalf of that user. Live Mode can surface:
- Private pages and databases that the user can access in Notion.
- Newly created or recently updated content that has not yet been picked up by the scheduled crawl.
Permissions & Security
The security behavior and data handling differ significantly between the two retrieval modes:Indexed Content (Data Crawling)
- Permission Propagation: Notion’s crawl APIs do not expose full user/group-level sharing lists. Consequently, Glean cannot project document ACLs into its index or enforce fine-grained permissions.
- Visibility: All indexed content shared with the integration is visible to all Glean users who can access the Notion connector.
- Compliance Recommendation: Glean recommends sharing only Notion content that is appropriate for organization-wide visibility.
Live Mode (MCP-based Fetching)
- Permission Behavior: Uses per-user OAuth tokens to call Notion’s MCP interface at query time. Notion enforces its own permission model, ensuring users only see content they can access directly in Notion.
- Data Handling: Queries fetch content on demand. Glean does not replace the indexed corpus with Live Mode responses or expand visibility beyond a user’s native access.
- Security Standards: Standard Glean security practices for data in transit and at rest apply, consistent with other connectors using activity-based updates.
Limitations
- ACL & Group Visibility: Notion’s API does not provide visibility into group membership or document-level access lists. It does not expose document-level permission lists in a way Glean can use to project per-user ACLs.
- Permission Model Scope: Glean does not support indexing documents not explicitly shared with the integration or those residing outside the permissions model allowed by the Notion API.
- Object-Level Propagation: There is no support for Notion object-level user permissions propagation for the indexed corpus.
- Content Redlisting: There is no mechanism to “redlist” or exclude specific child page content from the integration once a parent is shared.
- Unsupported Data: Glean does not yet support databases with multiple data sources. Embedded content types and external file attachments may not be indexed in the same way as native text.
- No Webhooks: Glean does not support webhooks for Notion; updates are captured only during scheduled crawls.
Requirements
This connector has specific technical and operational requirements that must be met before it can be configured.Technical Requirements
- You must have Glean Admin access to install and configure the connector.
- You must have Notion Workspace Admin privileges to create and manage integrations.
- A supported Notion Enterprise or Business plan with API access enabled.
- Use a dedicated admin or service account to create the integration. This prevents service disruption if an individual admin leaves the organization.
Credential Requirements
- A Bearer token generated via Notion My-Integrations. This token is required for all API calls related to data crawling and indexing.
- Required specifically for Live Mode. These tokens are issued to individual users when they connect their Notion accounts within Glean. They are managed separately from the workspace-level integration token.
Permission and Content Scope
- The Notion API requires that pages and databases be explicitly shared with the Glean integration app to be indexed.
- Because the indexing API does not support per-user ACLs, all shared content is visible to all Glean users with access to the connector.
- Admins must ensure that only content appropriate for organization-wide visibility is shared with the integration.
Live Mode Requirements
To enable real-time, permission-aware data fetching via Live Mode, the following additional conditions must be met:- A Glean admin must toggle the “Live Mode” (data fetching) option in the Glean Admin Console. If this option is not visible, contact your Glean representative to confirm eligibility.
- Individual users must complete the OAuth flow to link their Notion accounts to Glean.
- Live Mode may be restricted for environments with strict connector-level crawl restrictions or Data Loss Prevention (DLP) policies that conflict with real-time fetching behavior.
Configuration and Setup Instructions
Proper setup of the Notion connector happens in two places:- In Notion, where you create and configure the “Glean” internal integration.
- In the Glean Admin console, where you add the Notion data source, enter credentials, choose data retrieval modes (indexing and Live Mode), and start indexing pages.
Step 1: Create the Glean internal integration in Notion
-
Create a new integration by going to https://www.notion.com/my-integrations.

- Select the New integration button. Name the integration “Glean”.
-
Associate the integration with your workspace by selecting it from the drop down menu. (You can only do this if you are an admin of that workspace.) Select Submit.

Step 2: Configure the integration
-
Select Capabilities and select capabilities:

- Select Save changes.
- Select Distribution and ensure that the integration is not public.
Step 3: Add and configure the Notion data source in Glean
- In the Glean Admin console, go to Data sources.
- Click Add data source and select Notion.
- On the Notion connector setup page:
- Name and icon
- Enter a Name for this data source (for example,
Notion – Main workspace). - (Optional) Choose or upload an Icon. This label and icon will appear in Glean search results and filters.
- Enter a Name for this data source (for example,
- Choose your retrieval modes:
- Data crawling and indexing (recommended): Enable this to allow Glean to crawl Notion via the workspace integration token and build a shared Notion index. Glean will index pages and databases that you explicitly share with the integration.
- Data fetching (Live Mode) (optional): Enable this to allow Glean to use per‑user OAuth and MCP to fetch fresh, permission‑aware Notion content at query time.
- Credentials
In the credentials section, paste the values you copied from Notion:- Internal integration token – paste the token from the Notion integration’s Secrets/API view.
- Workspace – paste or select the corresponding workspace value (workspace name or ID as displayed in Notion).
- Verify there are no extra spaces before or after the token or workspace values.
- Name and icon
- Click Save to validate the connection.
Glean will verify it can reach the Notion API using the integration token you provided.
- The token is from the correct integration and workspace.
- The integration is still enabled in Notion and has the required read capabilities.
- The token and workspace values are copied exactly as shown in Notion.
Step 4: (Optional) Enable per‑user Notion OAuth for Live Mode
If you enabled Live Mode in Step 3, users must authorize Glean to act on their behalf in Notion.- After the admin enables Live Mode:
- Users will see a Connect Notion prompt in Glean (for example in Glean Assistant or search).
- Each user who wants permission‑aware Notion answers should:
- Click Connect Notion in Glean.
- Review and accept the Notion OAuth consent screen when prompted.
- Complete the flow to return to Glean.
- Once connected:
- Glean Assistant can call Notion via MCP at query time using that user’s OAuth token.
- These Live Mode answers are combined with the existing indexed corpus at answer time; they do not change the underlying index or expand access for other users.
Step 5: Share pages and databases with the Glean integration (for indexing)
For the indexed corpus (Data crawling and indexing) to include a page or database, it must be explicitly shared with the Glean integration in Notion.- In Notion, navigate to a page (or database) you want Glean to index.
- Open the page menu (three‑dot menu in the top‑right corner).
- Select Connections > Add connections> Glean (or select the name of the integration you created).
- Repeat for any additional pages or top‑level spaces you want to include. Child pages of a shared page are automatically shared with the integration.
- Review your sharing choices carefully:
- For the indexed corpus, any content shared with the Glean integration becomes visible in Glean to all users who can access the Notion connector.
- Only share pages and databases here that you are comfortable making broadly available within your company.