Integrate with Notion for Data Operations

The Notion Tool lets an agent load and update documents in Notion. It exposes three functions: load_data (loads a list of pages or databases by ID), search_data (searches for matching pages or databases), and append_data (appends content to a matching page or database). The user provides an API token to initialize NotionToolSpec.

Use it when an agent needs to read from or write to Notion -- for example appending a heading to an existing page, as in agent.run("Append the heading 'I am legend' to the movies page"). It requires a Notion API token, so it is not usable without first setting one up in your Notion workspace.

Given a page or database ID, load_data returns its content; a search query, search_data returns matching pages or databases; and content plus a target page, append_data writes it in. Attach it to an agent via to_tool_list():

from llama_index.tools.notion import NotionToolSpec
from llama_index.core.agent.workflow import FunctionAgent
from llama_index.llms.openai import OpenAI

tool_spec = NotionToolSpec()

agent = FunctionAgent(
    tools=tool_spec.to_tool_list(),
    llm=OpenAI(model="gpt-4.1"),
)

print(await agent.run("Append the heading 'I am legend' to the movies page"))

Developers building LlamaIndex agents that need to read from, search, or write to a Notion workspace as part of a larger workflow.