Workflow Templates

What is this workflow? This n8n template automates the process of adding an AI-generated summary at the top of your WordPress posts. It retrieves, processes, and updates your posts dynamically, ensuring efficiency and flexibility without relying on a heavy WordPress plugin. Example of AI Summary Section How It Works Triggers → Runs on a scheduled interval or via a webhook when a new post is published. Retrieves posts → Fetches content from WordPress and converts HTML to Markdown for AI processing. AI Summary Generation → Uses OpenAI to create a concise summary. Post Update → Inserts the summary at the top of the post while keeping the original excerpt intact. Data Logging & Notifications → Saves processed posts to Google Sheets and notifies a Slack channel. Why use this workflow? ✅ No need for a WordPress plugin → Keeps your site lightweight. ✅ Highly flexible → Easily connect with Google Sheets, Slack, or other services. ✅ Customizable → Adapt AI prompts, formatting, and integrations to your needs. ✅ Smart filtering → Ensures posts are not reprocessed unnecessarily. 💡 Check the detailed sticky notes for setup instructions and customization options!

This workflow automates the creation of exam questions (both open-ended and multiple-choice) from educational content stored in Google Docs, using AI-powered analysis and vector database retrieval This workflow saves educators hours of manual work while ensuring high-quality, curriculum-aligned assessments. Let me know if you'd like help adapting it for specific subjects! Use Cases Educators**: Rapidly generate quizzes, midterms, or flashcards. E-learning platforms**: Automate question banks for courses. Corporate training**: Create assessments for employee onboarding. Technical Requirements: APIs**: Google Gemini, OpenAI, Qdrant, Google Workspace. n8n Nodes**: LangChain, Google Sheets/Docs, HTTP requests, code blocks. This workflow combines AI efficiency with human-curated quality, making it a powerful tool for modern education and training. Advantages of This Workflow ✅ Fully Automated Exam Generation: From document to fully formatted quiz content with no manual intervention. ✅ Supports Comprehension and Critical Thinking: Questions are designed to go beyond factual recall, including inference and application. ✅ Uses AI and RAG for Accuracy: Ensures that answers are grounded in the document content, reducing hallucination. ✅ Seamless Google Integration: Pulls content from Google Docs and writes outputs to Google Sheets. ✅ Scalable for Any Subject: Works with any article or content domain as input. ✅ Modular and Customizable: Can be easily adapted to generate different question types or to use other LLMs or storage systems. How It Works Document Ingestion: The workflow starts by fetching an educational document (e.g., textbook chapter, lecture notes) from Google Docs. Converts the document to Markdown for structured processing. AI Processing: Splits text into chunks and generates vector embeddings (via OpenAI) for semantic analysis. Stores embeddings in Qdrant (vector database) for retrieval. Question Generation: Open-ended questions: Google Gemini AI creates 10 critical-thinking questions. Multiple-choice questions: Generates 10 MCQs (1 correct + 3 plausible distractors) using RAG to validate answers against the vector DB. Answer Validation: For open questions: Retrieves context-aware answers from the vector store. For MCQs: Ensures distractors are incorrect but believable via AI cross-checking. Output: Saves questions/answers to Google Sheets in two tabs: Open questions: Question + AI-generated answer. Closed questions: MCQ + options + correct answer. Set Up Steps Prerequisites: APIs/Accounts: Google Workspace (Docs + Sheets). OpenAI (for embeddings). Google Gemini (for question generation). Qdrant (vector DB – self-hosted or cloud). n8n Nodes: Ensure LangChain, Google Sheets/Docs, and HTTP request nodes are installed. Configure Connections: Link credentials for: Google Docs/Sheets (OAuth2). OpenAI (API key). Google Gemini (API key). Qdrant (URL + API key). Customize Input: Replace the default Google Doc ID in the "Get Doc" node with your source document. Adjust chunk size/overlap (Token Splitter node) for optimal text processing. Tweak Question Generation: Modify prompts in: "Open questions" node: Adjust criteria (e.g., difficulty, question types). "Closed questions" node: Edit MCQ formatting rules. Output Settings: Update the Google Sheet ID in "Write open" and "Write closed" nodes. Map columns in Google Sheets to match question/answer formats. Run & Automate: Trigger manually ("Test workflow") or schedule periodic runs (e.g., for updated content). Need help customizing? Contact me for consulting and support or add me on Linkedin.

How It Works: This n8n template automates the process of tracking design changes in Figma and updating relevant Jira issues. The template is triggered when a new version is created in Figma via a custom plugin. Once the version is committed, the plugin sends the design details to an n8n workflow using a webhook. The workflow then performs the following actions: Fetches the Jira issue based on the provided issue link from Figma. Adds the design changes as a comment to the Jira issue. Updates the status of the Jira issue based on the provided task status (e.g., "In Progress", "Done"). This streamlines the workflow, reducing the need for manual updates and ensuring that both the design team and developers have the latest design changes and task statuses in sync. How to Use It: Set up the Figma Plugin: Install the Figma Commit Plugin from GitHub. In the plugin, fill out the version name, design link, Jira issue link, and the task status. Commit the changes in Figma, which will trigger the webhook. Set Up the n8n Workflow: Import this template into your n8n instance. Connect the Figma Trigger node to capture version updates from Figma. Configure the Jira nodes to retrieve the issue and update the status/comment based on the data sent from the plugin. Automate: Once the version is committed in Figma, the workflow will automatically update the Jira issue and keep both your Figma design and Jira tasks in sync! By integrating Figma, Jira, and n8n through this template, you’ll eliminate manual steps, making collaboration between design and development teams more efficient.

🚀 LinkedIn Comments to Leads Extractor & Enricher (Apify) → Google Sheets / CSV Overview Automate LinkedIn lead generation by scraping comments from targeted posts and enriching profiles with detailed data This n8n workflow automatically extracts leads from LinkedIn post comments using Apify's powerful scrapers (no LinkedIn login required), enriches the data with additional profile information, and exports everything to Google Sheets or CSV format. ✨ Key Features 🔍 No Login Required: Scrape LinkedIn data without sharing credentials 💰 Cost-Effective: First 1,000 comments are free with Apify 📊 Data Enrichment: Enhance basic comment data with full profile details 📈 Export Options: Choose between Google Sheets or CSV output 🎯 Targeted Scraping: Focus on specific posts for quality leads 🛠️ Apify Scrapers Used 1. LinkedIn Post Comments Scraper Tool**: LinkedIn Post Comments, Replies, Engagements Scraper | No Cookies Pricing**: $5.00 per 1,000 results Function**: Extracts all comments and engagement data from specified LinkedIn posts 2. LinkedIn Profile Batch Scraper Tool**: LinkedIn Profile Details Batch Scraper (No Cookies Required) Pricing**: $5.00 per 1,000 results Function**: Enriches scraped profiles with detailed information > 💡 Free Tier: Apify provides 1,000 free scraped comments to get you started! 📋 Prerequisites Required API Credentials Apify Token Add your APIFY_TOKEN to the workflow credentials Get your token from Apify Console Google Sheets Credentials (if using Sheets export) Configure OAuth credentials for Google Sheets integration Follow n8n's Google Sheets setup guide 🔄 Workflow Process Default Mode: Form-Based Execution Manual Trigger → Launches the workflow Form Submission → User-friendly form for inputting LinkedIn post URLs Comment Scraping → Apify extracts all comments from specified posts Profile Enrichment → Additional profile data gathered for each commenter Data Processing → Creates unique, enriched lead list Google Sheets Export → Automatically populates your spreadsheet Result: You'll be redirected to a Google Sheets document containing all enriched leads Alternative Mode: CSV Export For users preferring CSV output: Disable: Form trigger nodes Enable: Manual trigger node Disable: Google Sheets export nodes Enable: CSV download nodes Configure: Add post IDs/URLs in "Set manual fields" node Execute: Run workflow and download CSV from the CSV node 📊 Output Data Structure Your exported data will include: Basic Info**: Name, headline, location Profile Details**: Company, position, industry Engagement Data**: Comment content, engagement metrics Contact Info**: Available profile links and connections Enriched Data**: Additional profile insights from Apify 💡 Pro Tips Quality over Quantity**: Target posts with high-quality, relevant engagement Monitor Costs**: Track your Apify usage to stay within budget Data Hygiene**: Regularly clean and deduplicate your lead lists Compliance**: Ensure your scraping activities comply with LinkedIn's terms of service 🆘 Troubleshooting Common Issues: Authentication Errors**: Verify your Apify token is correctly configured Empty Results**: Check that your LinkedIn post URLs are valid and public Export Failures**: Ensure Google Sheets credentials are properly set up Need Help? Contact Saverflow.ai for support and custom workflow development.

Capture leads from Tally forms to Google Sheets CRM This workflow automates lead intake by capturing form submissions, enriching them with smart tags and scores, storing them in a Google Sheets CRM, and sending personalized welcome emails. Who's it for This template is perfect for solopreneurs, small agencies, and marketing consultants who need a simple yet powerful CRM system without paying for expensive tools like HubSpot or Salesforce. If you're collecting leads through forms and manually copying data to spreadsheets, this automation will save you hours every week. What it does When someone submits your Tally lead capture form, this workflow automatically: Generates a unique lead ID and timestamps the entry Analyzes their responses to assign relevant service tags (Podcast Lead, Social Content Lead, Video Editing Lead, etc.) Calculates an initial lead score based on interest level Determines the next follow-up date automatically Appends all data to your Google Sheets CRM with proper formatting Logs the activity in a separate tracker sheet Sends a personalized welcome email tailored to their interests Updates the lead status to "Nurturing" Requirements Apps & Services: Google Sheets (for your CRM database) Tally.so or Google Forms (for lead capture) SendGrid account (free tier works) for email sending n8n instance (cloud or self-hosted) Setup needed: Create the Google Sheets CRM structure (4 sheets: Leads, Email Sequence Tracker, Activity Log, Dashboard Data) Set up your Tally form with fields: Name, Email, Business Type, Interest Level, Services Needed Configure SendGrid API credentials in n8n Connect your Google Sheets account to n8n How to set up Copy the Google Sheets template with all four sheets (Leads, Email Sequence Tracker, Activity Log, Dashboard Data) and set up column headers as specified Create your Tally form at tally.so with the required fields, then grab the webhook URL from n8n Import this workflow into your n8n instance Configure credentials for Google Sheets and SendGrid Update the webhook URL in your Tally form settings Customize the welcome email in the "Generate Welcome Email" node with your branding Test with a sample submission to verify everything flows correctly Activate the workflow and start capturing leads How to customize Modify service tags: Edit the serviceTagMap object in the "Auto-Tag Lead" node to match your specific services. Adjust lead scoring: Change the scoring logic in "Calculate Initial Dates & Score" to prioritize different interest levels or sources. Personalize emails: Update the email templates in "Generate Welcome Email" to match your brand voice and add specific resources or links. Add more integrations: Extend the workflow with Slack notifications, SMS alerts via Twilio, or sync to other tools like Notion or Airtable. Change follow-up timing: Modify the daysToAdd calculation to adjust when leads receive their next touchpoint. Add conditional paths: Use IF nodes to route different lead types to different email sequences or team members.

This n8n template from Intuz provides a complete solution to automate your accounting by instantly creating QuickBooks sales receipts for every new Stripe payment. This workflow automates the process of recording successful payments from Stripe into QuickBooks by creating corresponding Sales Receipts. It ensures payment data is captured accurately, checks whether the customer exists in QuickBooks, and creates a new customer if necessary before generating the receipt. This integration streamlines bookkeeping by eliminating manual data entry and ensuring all payment records are synchronized between systems. Who's this workflow for? Accountants & Bookkeepers Small Business Owners E-commerce Managers Finance Teams How it works 1. Trigger on Successful Payment: The workflow starts instantly when a payment_intent.succeeded event is received from Stripe via a webhook. This means it only runs after a payment is confirmed. 2. Get Customer Details: It uses the customer ID from the payment to fetch the customer's full details (name and email) from Stripe. 3. Check for Customer in QuickBooks: The workflow then searches your QuickBooks account to see if a customer with that name already exists. 4. Create Customer if New: If the customer is not found in QuickBooks, a new customer record is automatically created using the information from Stripe. 5. Generate Sales Receipt: Finally, using the correct customer record (either existing or newly created) and the payment amount, the workflow creates and saves a new sales receipt in QuickBooks, perfectly matching the Stripe transaction. Key Requirements to Use This Template 1. n8n Instance: An active n8n account (Cloud or self-hosted). 2. Stripe Account: An active Stripe account with API access. You must be able to create and manage webhooks. 3. QuickBooks Online Account: An active QuickBooks Online account with API access to manage customers and sales receipts. Setup Instructions 1. Configure the Webhook Trigger: Copy the webhook URL from the Capture Payment (Webhook) node in n8n. In your Stripe dashboard, go to Developers > Webhooks and add a new endpoint. Paste the n8n webhook URL and have it listen for the payment_intent.succeeded event. 2. Connect Stripe: In the Get a customer node, connect your Stripe account credentials. 3. Connect QuickBooks: In all three QuickBooks nodes (Find Customer, Create a customer, and Create a payment), connect your QuickBooks Online account using OAuth2 credentials. 4. Activate Workflow: Save the workflow and toggle the "Active" switch to ON. Your accounting automation is now live! Connect with us Website: https://www.intuz.com/services Email: getstarted@intuz.com LinkedIn: https://www.linkedin.com/company/intuz Get Started: https://n8n.partnerlinks.io/intuz For Custom Worflow Automation Click here- Get Started

This workflow automates the process of estimating a person’s fashion size from an uploaded image using an AI model. This workflow is an automated pipeline that uses an AI model to estimate a person's body measurements and clothing size from an image URL. Key Features 🔁 Full Automation** – From image submission to result display, the process requires no manual steps. ⚙️ Easy Integration** – Uses n8n’s native nodes and simple HTTP requests to connect with Fal.ai’s API. 🕒 Real-Time Processing** – Automatically waits and checks for the AI result, ensuring the user receives the output as soon as it’s ready. 🧩 Modular Design** – Each step (submit → process → check → result) is clearly separated, making it easy to modify or extend (e.g., adding notifications or storing results in a database). 💡 User-Friendly Interface** – The initial form and final result form make it accessible even for non-technical users. 🔐 Secure** – Authentication to the Fal.ai API is handled through HTTP header authorization, keeping API keys protected. How it works Form Trigger: The workflow starts with a public form where a user submits a URL of an image. AI Processing Request: The submitted image URL is sent to the fal.run AI service (specifically, the "fashion-size-estimator" model) via a POST request. This initial request places the job in a queue and returns a unique request_id. Polling for Completion: The AI processing is asynchronous and takes some time. The workflow enters a loop where it: Waits: Pauses for 10 seconds to give the AI model time to process the request. Checks Status: Uses the request_id to check the status of the job. Conditional Check: An IF node checks if the status is "COMPLETED". If NO (not completed), the loop repeats (wait, then check again). If YES, the workflow exits the loop. Fetching and Displaying Results: Once processing is complete, the workflow retrieves the final result (containing the size, height, bust, waist, and hip measurements) and automatically displays it to the user on a "thank you" page. Set up steps To make this workflow operational, you need to configure the API authentication. Obtain an API Key: Create an account at fal.ai Navigate to your account settings to generate an API key. Configure Credentials in n8n: In your n8n instance, create a new HTTP Header Auth credential (you can name it "Fal.run API"). Set the Name field to Authorization. Set the Value field to Key YOURAPIKEY, replacing "YOURAPIKEY" with the actual key you obtained from fal.ai. Ensure this credential is correctly selected in the three HTTP Request nodes: "Send image to estimator", "Get status", and "Get result". Need help customizing? Contact me for consulting and support or add me on Linkedin.

📄 What this workflow does This workflow transforms your n8n instance into a fully automated AI sales assistant for WooCommerce stores. It detects customer intent from chat, searches products, answers FAQs, generates Stripe payment links, captures leads into your CRM, and even escalates to human support when needed. It provides smooth conversational memory and syncs with your knowledge base to ensure accurate, human-like responses. 👤 Who is this for WooCommerce store owners who want to automate customer support and sales. Sales and marketing teams looking to scale personalized product recommendations. E-commerce managers who want to reduce manual chat handling. Anyone aiming to integrate AI assistants with payments, CRM, and FAQs. ✅ Requirements WooCommerce account with API access. Qdrant vector store (for FAQ and RAG retrieval). OpenAI/Gemini API credentials (for intent detection + message generation). Google Drive account (to sync and update knowledge base docs). Stripe account (to generate instant payment links). CRM account (HubSpot, Pipedrive, etc.) if lead capture is required. Telegram account for optional human escalation. ⚙️ How to set up Connect WooCommerce API credentials in n8n. Configure Gemini/OpenAI API for intent extraction and chat generation. Set up Qdrant for document retrieval, and link to your Google Drive Sales Docs folder. Configure Stripe API to enable instant payment link generation. Connect your CRM to capture new leads automatically. Add Telegram bot credentials for human escalation (optional). Enable conversational memory and test chat flows end-to-end. 🔁 How it works Intent Extraction → AI analyzes chat messages to detect Product Search, FAQ, Payment, or Lead Capture. Product Search → Queries WooCommerce catalog by keyword, SKU, or price range. FAQ Answering → Retrieves company policies/docs from Qdrant + Google Drive, answered via RAG. Payment Links → Stripe generates instant checkout links for customers ready to buy. Lead Capture → Name + email are auto-stored into CRM. Human Escalation → If intent is unclear, conversation is forwarded to Telegram. Conversational Memory → Maintains last 12 exchanges for natural dialogue. Knowledge Sync → Google Drive docs auto-update into Qdrant for live FAQ support. 💡 About Margin AI Margin AI is an AI-services agency that acts as your AI Service Companion. We design intelligent, human-centric automation solutions—turning your team’s best practices into scalable workflows and tools. Industries like marketing, sales, and operations benefit from our tailored AI consulting, automation tools, and chatbot development.

Automated WhatsApp Welcome Messages for Sales Leads with Google Sheets & Rapiwa Who is this for? This automation is ideal for sales teams, digital marketers, support agents, or small business owners who collect leads in Google Sheets and want to automatically send WhatsApp welcome messages. It's a cost-effective and easy-to-use solution built for those not using the official WhatsApp Business API but still looking to scale communication. What this Workflow Does This n8n automation reads leads from a connected Google Sheet, verifies if the provided WhatsApp numbers are valid using the Rapiwa API, and sends a personalized welcome message. It updates the sheet based on delivery success or failure, and continues this process every 5 minutes — ensuring new leads are automatically engaged. Key Features Automatic Scheduling**: Runs every 5 minutes (adjustable) Google Sheets Integration**: Reads and updates lead data WhatsApp Number Validation**: Confirms number validity via Rapiwa Personalized Messaging**: Uses lead name for custom messages Batch Processing**: Sends up to 60 messages per cycle Safe API Usage**: Adds 5-second delay between each message Error Handling**: Marks failed messages as not sent and unverified Live Status Updates**: Sheet columns are updated after each attempt Loop Logic**: Repeats continuously to catch new rows How to Use Step-by-step Setup Prepare Your Google Sheet Copy this Sample Sheet Ensure it includes the following columns: WhatsApp No name (note: trailing space is required) row_number status, check, validity Connect Google Sheets in n8n Use OAuth2 credentials to allow n8n access Set the workflow to fetch rows where check is not empty Get a Rapiwa Account Sign up at https://rapiwa.com Add your WhatsApp number Retrieve your Bearer Token from your Rapiwa dashboard Configure HTTP Request Nodes Use Rapiwa's API endpoints: Verify Number: https://app.rapiwa.com/api/verify-whatsapp Send Message: https://app.rapiwa.com/api/send-message Add your Bearer Token to the header Start Your Workflow Run the n8n automation It will read leads, clean phone numbers, verify WhatsApp validity, send messages, and update the sheet accordingly Requirements A Google Sheet with correctly formatted columns Active Rapiwa subscription (~$5/month) A valid Bearer Token from Rapiwa Your WhatsApp number connected to Rapiwa n8n instance with: Google Sheets integration (OAuth2 setup) HTTP Request capability Google Sheet Column Reference | name | number | email | time | check | validity | status | |-----------------|--------------|-------------------|-----------------------------|---------|------------|-----------| | Abdul Mannan | 8801322827799| contact@spagreen.net| September 14th 2025, 10:34 | checked | verified | sent | | Abdul Mannan | 8801322827798| contact@spagreen.net| September 14th 2025, 10:34 | checked | unverified | not sent | Workflow Logic Summary Trigger Every 5 Minutes Fetch All Rows with Pending Status Limit to 60 Rows per Execution Clean and Format Phone Numbers Check Number Validity via Rapiwa Condition Check: If valid → Send Message If invalid → Update status as not sent, unverified Send WhatsApp Message via Rapiwa Update Sheet Row On success: sent, verified, checked On failure: not sent, unverified Delay 5 seconds before next message Repeat for next lead Customization Ideas Add image or document sending support via Rapiwa Customize messages based on additional fields (e.g., product, service) Log failures to a separate sheet Send admin email for failed batches Add support for multilingual messages Notes & Warnings The column name "name " includes a space — do not remove or rename it. International number format is required for Rapiwa to work correctly. If you're sending many messages, increase the Wait node delay to prevent API throttling. Support WhatsApp Support: Chat Now Discord: Join SpaGreen Community Facebook Group: SpaGreen Support Website: https://spagreen.net Developer Portfolio: Codecanyon SpaGreen

📍Overview This no-code workflow is built for creators, agencies, and operators who want to automate the repurposing of Instagram Reels. It runs end-to-end and outputs structured insights and content-ready scripts—without touching a single tool manually. 🧰 What It Does Triggered simply by sending an Instagram Reel URL via Telegram. Downloads the Reel automatically. Converts video to audio using FreeConvert API. Transcribes speech to text using AssemblyAI. Analyzes both transcript and description using a connected LLM (OpenAI or Mistral). Extracts: Niche Core message 3 viral content hooks 3 ready-to-use short-form video scripts Saves all data to a Google Sheet for easy reuse by the creator or team. 🧪 APIs & Integrations Telegram Bot API (for triggering) FreeConvert API (MP4 to MP3 conversion) AssemblyAI (for transcription) OpenAI or Mistral (LLM for content analysis) Google Sheets API (for logging all outputs) ✅ Requirements An n8n instance (self-hosted or cloud) AssemblyAI API key FreeConvert API key Telegram Bot token Google service account credentials Your preferred LLM key (OpenAI or Mistral) 💡 Why Use This Workflow Runs entirely from Telegram—no dashboards required Helps you extract deep insights and reusable content from any Instagram Reel All tools used are free or very low cost Ideal for scaling personal brands or agency operations

Overview When designing user interfaces, toolbar icons often get overlooked, even though their placement and grouping dramatically impact usability and user flow. This workflow leverages Gemini AI to automatically analyze UI screens, classify toolbar icons based on Apple’s Human Interface Guidelines (HIG), and suggest optimal placements. By combining AI analysis with structured placement logic, this workflow helps designers build more consistent, efficient, and user-friendly interfaces—without spending hours manually arranging icons. 🚀 Features AI Classification**: Uses Gemini AI to analyze screenshots and classify icons into roles like .primaryAction, .navigation, .confirmationAction, and more. HIG-Based Placement**: Automatically assigns icons to the correct toolbar areas—Leading (Left), Trailing (Right), Center, Bottom, or System-decided. Usage-Aware Reordering**: Reorders icons based on frequency of use so the most relevant actions appear where users expect them. JSON Output**: Delivers structured results for seamless integration into design tools or documentation. 🔧 Setup Instructions Install the Workflow: Import the workflow into your n8n instance. Configure Input: Upload a screenshot of your UI. Upload a set of icons you want to classify and place. Set Up Gemini AI Node: Add your Gemini AI API key in the node’s credentials. Run the Workflow: Submit the inputs and let the AI classify and assign placements. Export Results: Copy the JSON output or connect the workflow to your preferred design/documentation tools. ⚙️ How It Works Form Submission – Capture screenshot + icons. Gemini AI Agent – Interprets screen context and classifies each icon. Placement Logic – Maps icons to the correct toolbar areas. Reordering – Adjusts order based on relevance and HIG standards. Structured Output – Produces clean JSON for further use. 🎨 Customization Change AI Prompts**: Modify the Gemini AI node prompts to reflect your app’s design language. Adjust Placement Rules**: Update logic to follow custom guidelines beyond Apple HIG. Integrate with Design Tools**: Send the JSON output directly to tools like Figma, Sketch, or internal systems. 💡 Why This Matters Consistency**: Ensures toolbar designs always follow Apple’s HIG. Efficiency**: Saves designers hours of manual icon placement. Scalability**: Works across multiple screens, flows, and apps. AI-Assisted Design**: Augments designer decisions with structured insights instead of replacing them.

This workflow implements a Retrieval-Augmented Generation (RAG) system that integrates Google Drive and Qdrant. This setup creates a powerful, self-updating knowledge base that provides accurate, context-aware answers to user queries. Key Advantages Automated Knowledge Base Updates** No manual intervention is required—documents in Google Drive are automatically synchronized with Qdrant. Efficient Search and Retrieval** Vector embeddings enable fast and precise retrieval of relevant information. Scalable and Flexible** Works with multiple documents and supports continuous growth of your dataset. Seamless AI Integration** Combines OpenAI embeddings for vectorization and Google Gemini for high-quality natural language answers. Metadata-Enhanced Storage** Each document stores metadata (file ID and name), making it easy to manage and track document versions. End-to-End RAG Pipeline** From document ingestion to AI-powered Q\&A, everything is handled inside one n8n workflow. How It Works This workflow implements a Retrieval-Augmented Generation (RAG) system that automatically processes, stores, and retrieves document information for AI-powered question answering. Here’s how it functions: Document Processing & Vectorization: The system monitors a specified Google Drive folder for new or updated files. When a file is added or modified, it is downloaded and split into manageable chunks using a Recursive Character Text Splitter. Each chunk is converted into vector embeddings using OpenAI's embedding model. These vectors, along with metadata (file ID, file name), are stored in a Qdrant vector database. Automatic Updates: The workflow includes a mechanism to delete old vectors associated with an updated file before inserting the new ones, ensuring the knowledge base remains current. Query Handling & Response Generation: When a user sends a chat message (via a chat trigger), the system: Retrieves the most relevant document chunks from Qdrant based on the query's semantic similarity. Uses a Google Gemini language model to generate a context-aware answer grounded in the retrieved documents. This provides accurate, source-based responses instead of relying solely on the AI's internal knowledge. Initial Setup & Maintenance: The workflow can be triggered manually to create the Qdrant collection or clear all existing data. It processes all existing files in the Drive folder during initial setup, populating the vector store. Set Up Steps To configure this workflow, follow these steps: STEP 1: Create Qdrant Collection Replace QDRANTURL in the "Create collection" and "Clear collection" nodes with your Qdrant instance URL (e.g., http://your-qdrant-host:6333). Replace COLLECTION with your desired collection name. Ensure the Qdrant API credentials are correctly set in the respective HTTP Request nodes. STEP 2: Configure Google Drive Access Set up OAuth credentials for Google Drive to allow the workflow to: Read files from a specific folder . Download files for processing. Update the Folder ID in the "Search files" and "Update?" trigger nodes to point to your target Google Drive folder. STEP 3: Set Up AI Models Configure the OpenAI API credentials in the Embeddings nodes for generating text embeddings. Configure the Google Gemini (PaLM) API credentials in the Google Gemini Chat Model node for generating answers. STEP 4: Configure Metadata The system automatically attaches metadata (file_id, file_name) to each document chunk. This is set in the Default Data Loader nodes. This metadata is crucial for identifying the source of information and for the update mechanism. STEP 5: Test the RAG System The workflow includes a chat trigger ("When chat message received") for testing. Send a query to test the retrieval and answer generation process. Need help customizing? Contact me for consulting and support or add me on Linkedin.