Prod: Notion to Vector Store - Dimension 768

For Notion, this automated workflow efficiently transfers newly added page content into a vector store, filtering out non-text elements and summarizing the information for enhanced data retrieval. It leverages LangChain for embedding and organizes content for improved accessibility, ensuring timely updates every minute.

7/8/2025
8 nodes
Medium
vs70y1mj5s2xzuapmanualmediumlangchainnotiontriggernotionfiltersummarize
Categories:
Manual TriggeredMedium Workflow
Integrations:
LangChainNotionTriggerNotionFilterSummarize

Target Audience

This workflow is ideal for:
- Content Creators: Those who frequently add new pages to Notion and want to automatically process and store content efficiently.
- Data Analysts: Professionals needing to filter and summarize text data from Notion for insights.
- Developers: Individuals looking to integrate Notion with vector databases for better data retrieval and analysis.
- Businesses: Organizations that utilize Notion for documentation and want to enhance their data management capabilities.

Problem Solved

This workflow addresses the challenge of managing and processing new content added to Notion. It automates the extraction, filtering, and storage of relevant text data, ensuring that non-text content (like images and videos) is excluded. The workflow also summarizes the content and stores it in a vector database, making it easier to search and retrieve information later.

Workflow Steps

  • Trigger: The workflow is manually triggered when a new page is added to a specified Notion database.
    2. Retrieve Content: It retrieves all content from the newly added Notion page.
    3. Filter Content: The workflow filters out non-text content, ensuring only relevant text data is processed.
    4. Summarize: The remaining text content is concatenated into a single summary for easier handling.
    5. Create Metadata: Metadata such as pageId, createdTime, and pageTitle is created to accompany the content.
    6. Embed Content: The summarized content is embedded using Google Gemini for further processing.
    7. Store in Vector Database: Finally, the processed content and its metadata are inserted into a Pinecone vector store for efficient retrieval.

    • Customization Guide

    Users can customize this workflow by:
    - Changing Database ID: Update the databaseId in the Notion trigger to point to a different Notion database.
    - Modifying Filter Conditions: Adjust the filter conditions in the 'Filter Non-Text Content' node to include or exclude different types of content.
    - Customizing Summarization: Modify the summarization parameters in the 'Summarize - Concatenate Notion's blocks content' node to change how content is aggregated.
    - Adjusting Metadata Fields: Add or remove metadata fields in the 'Create metadata and load content' node to fit specific requirements.
    - Switching Vector Store: Change the Pinecone index in the 'Pinecone Vector Store' node if a different storage location or configuration is needed.