Notion

Text

Overview

The Notion connector is a text connector that fetches pages from Notion via the Notion API and extracts claims from page block content using db.ingest_text().

It supports querying a specific Notion database by ID, or performing a global workspace search across all pages the integration has access to. Each page is fetched individually, its blocks are read, and the plain text is assembled and fed into the text extraction pipeline.

How it works

Lists pages from a database (/databases/{id}/query) or searches the workspace (/search)
Extracts the page title from page properties (the property with type "title")
Fetches child blocks for each page and concatenates plain text from supported block types
Passes the assembled text to db.ingest_text() for claim extraction
Source ID is set to notion:{page_id} for provenance tracking

Supported block types

paragraph
heading_1, heading_2, heading_3
bulleted_list_item
numbered_list_item
to_do

The connector is rate-limited with a 0.1-second sleep between API calls to stay within Notion's rate limits. Results are paginated automatically up to max_pages.

Setup

Create a Notion integration at notion.so/my-integrations. Choose "Internal integration" for private workspace access.
Copy the Internal Integration Token. It starts with ntn_. This is the value you pass as token.
Share target pages or databases with the integration. Open the page or database in Notion, click the ... menu, select "Add connections", and choose your integration. The connector can only access pages that have been explicitly shared.
Get the database ID (optional). Open the database in Notion as a full page. The URL will look like https://www.notion.so/{workspace}/{database_id}?v=... — the 32-character hex string before the ? is the database ID.

Dependencies

The Notion connector requires the requests package:

$ pip install requests

Parameters

Parameter	Required	Default	Description
`api_key`	Yes	—	Notion integration token (`ntn_...`)
`database_id`	No	`None`	Specific database to query. If omitted, searches the entire workspace.
`max_pages`	No	`100`	Maximum number of pages to fetch
`save`	No	`False`	Encrypt and persist token to the token store

Note: In db.connect(), api_key is passed as token.

Examples

Basic — search all pages

conn = db.connect("notion", token="ntn_...")
result = conn.run(db)

Specific database

conn = db.connect("notion",
    token="ntn_...",
    database_id="abc123def456...",
    max_pages=50,
)
result = conn.run(db)

With token persistence

conn = db.connect("notion",
    token=os.environ["NOTION_TOKEN"],
    database_id="abc123...",
    save=True,
)
result = conn.run(db)