ManualTrigger Automate

For ManualTrigger Automate, streamline your workflow by manually triggering an automated process that integrates Sticky Note, LangChain, and Google Drive. Effortlessly fetch files, split them into manageable chunks, and engage in interactive chats to extract valuable insights and citations, enhancing your productivity and information retrieval.

7/8/2025
20 nodes
Complex
manualcomplexsticky notelangchaingoogle driveadvanced
Categories:
Complex WorkflowManual Triggered
Integrations:
Sticky NoteLangChainGoogle Drive

Target Audience

This workflow is ideal for:
- Data Scientists looking to automate the process of document analysis and citation extraction.
- Researchers who need to interact with large documents and retrieve specific information efficiently.
- Developers seeking to integrate AI chat capabilities into their applications using Google Drive and vector databases.
- Educators wanting to create interactive learning tools that leverage document-based queries.

Problem Solved

This workflow addresses the challenge of efficiently extracting relevant information from large documents. By integrating with Google Drive and utilizing LangChain's capabilities, it allows users to:
- Download documents from Google Drive.
- Split documents into manageable chunks for easier processing.
- Store and retrieve relevant chunks using a vector database (Pinecone).
- Answer queries based on the content of the documents, providing citations for references.

Workflow Steps

  • Trigger the Workflow: The workflow begins when the user clicks the "Execute Workflow" button.
    2. Set File URL: The workflow sets the URL of the file to be downloaded from Google Drive.
    3. Download File: The specified file is downloaded from Google Drive.
    4. Load Document: The downloaded file is loaded as a binary document for processing.
    5. Chunking: The document is split into smaller, manageable chunks using a recursive character text splitter.
    6. Embedding Creation: Each chunk is converted into embeddings using OpenAI's API for further processing.
    7. Store in Vector Database: The embeddings are stored in a Pinecone vector database for efficient retrieval.
    8. Chat Trigger: The workflow listens for incoming chat messages, which will trigger the query process.
    9. Query Processing: When a message is received, the workflow determines how many chunks to send for processing.
    10. Retrieve Relevant Chunks: The workflow fetches the top chunks that match the user's query from the vector database.
    11. Prepare Context: The relevant chunks are formatted for the AI model to understand.
    12. Answer Generation: The AI model generates an answer based on the context provided, including citations from the document.
    13. Compose Final Response: The final response is composed, including the answer and the citations for reference.

    • Customization Guide

    Users can customize this workflow by:
    - Changing the Document Source: Update the Google Drive file URL in the Set File URL node to point to a different document.
    - Modifying Chunk Size: Adjust the chunkSize and chunkOverlap parameters in the Recursive Character Text Splitter node to control how the document is split.
    - Adjusting the AI Model: Change the model used in the OpenAI Chat Model node to utilize different versions of OpenAI's models.
    - Customizing Queries: Modify the prompts and parameters in the Answer the query based on chunks node to tailor the responses based on specific needs.
    - Adding Additional Nodes: Integrate additional nodes to extend functionality, such as logging responses or sending notifications.