Workflow Templates

Description: This n8n workflow automates a Discord bot to fetch messages from a specified channel and send AI-generated responses in threads. It ensures smooth message processing and interaction, making it ideal for managing community discussions, customer support, or AI-based engagement. This workflow leverages Redis for memory persistence, ensuring that conversation history is maintained even if the workflow restarts, providing a seamless user experience. How It Works The bot listens for new messages in a specified Discord channel. It sends the messages to an AI model for response generation. The AI-generated reply is posted as a thread under the original message. The bot runs on an Ubuntu server and is managed using PM2 for uptime stability. The Discord bot (Python script) acts as the bridge, capturing messages from Discord and sending them to the n8n webhook. The n8n workflow then processes these messages, interacts with the AI model, and sends the AI's response back to Discord via the bot. Prerequisites to host Bot Sign up on Pella, which is a managed hosting service for Discord Bots. (Easy Setup) A Redis instance for memory persistence. Redis is an in-memory data structure store, used here to store and retrieve conversation history, ensuring that the AI can maintain context across multiple interactions. This is crucial for coherent and continuous conversations. Set Up Steps 1️⃣ Create a Discord Bot Go to the Discord Developer Portal. Click “New Application”, enter a name, and create it. Navigate to Bot > Reset Token, then copy the Bot Token. Enable Privileged Gateway Intents (Presence, Server Members, Message Content). Under OAuth2 > URL Generator, select bot scope and required permissions. Copy the generated URL, open it in a browser, select your server, and click Authorize. 2️⃣ Deploy the Bot on Pella Create a new folder discord-bot and navigate into it: Create and configure an .env file to store your bot token: Copy the code to .env: (You can copy the webhook URL from the n8n workflow) TOKEN=your-bot-token-here WEBHOOK_URL=https://your-domain.tld/webhook/getmessage Create file main.py copy the below code and save it: Copy this Bot script to main.py: import discord import requests import json import os from dotenv import load_dotenv Load environment variables from .env file load_dotenv() TOKEN = os.getenv("TOKEN") WEBHOOK_URL = os.getenv("WEBHOOK_URL") Bot Configuration LISTEN_CHANNELS = ["YOUR_CHANNEL_ID_1", "YOUR_CHANNEL_ID_2"] # Replace with your target channel IDs Intents setup intents = discord.Intents.default() intents.messages = True # Enable message event intents.guilds = True intents.message_content = True # Required to read messages client = discord.Client(intents=intents) @client.event async def on_ready(): print(f'Logged in as {client.user}') @client.event async def on_message(message): if message.author == client.user: return # Ignore bot's own messages if str(message.channel.id) in LISTEN_CHANNELS: try: fetched_message = await message.channel.fetch_message(message.id) # Ensure correct fetching payload = { "channel_id": str(fetched_message.channel.id), # Ensure it's string "chat_message": fetched_message.content, "timestamp": str(fetched_message.created_at), # Ensure proper formatting "message_id": str(fetched_message.id), # Ensure ID is a string "user_id": str(fetched_message.author.id) # Ensure user ID is also string } headers = {'Content-Type': 'application/json'} response = requests.post(WEBHOOK_URL, data=json.dumps(payload), headers=headers) if response.status_code == 200: print(f"Message sent successfully: {payload}") else: print(f"Failed to send message: {response.status_code}, Response: {response.text}") except Exception as e: print(f"Error fetching message: {e}") client.run(TOKEN) Create requirements.txt and copy: discord python-dotenv 3️⃣ Follow the video to set up the bot which will run 24/7 Tutorial - https://www.youtube.com/watch?v=rNnK3XlUtYU Note: Free Plan will expire after 24 hours, so please opt for the Paid Plan in Pella to keep your bot running. 4️⃣ n8n Workflow Configuration The n8n workflow consists of the following nodes: Get Discord Messages (Webhook):** This node acts as the entry point for messages from the Discord bot. It receives the channel_id, chat_message, timestamp, message_id, and user_id from Discord when a new message is posted in the configured channel. Its webhook path is /getmessage and it expects a POST request. Chat Agent (Langchain Agent):** This node processes the incoming Discord message (chat_message). It is configured as a conversational agent, integrating the language model and memory to generate an appropriate response. It also has a prompt to keep the reply concise, under 1800 characters. OpenAI -4o-mini (Langchain Language Model):** This node connects to the OpenAI API and uses the gpt-4o-mini-2024-07-18 model for generating AI responses. It is the core AI component of the workflow. Message History (Redis Chat Memory):** This node manages the conversation history using Redis. It stores and retrieves chat messages, ensuring the Chat Agent maintains context for each user based on their user_id. This is critical for coherent multi-turn conversations. Calculator (Langchain Tool):** This node provides a calculator tool that the AI agent can utilize if a mathematical calculation is required within the conversation. This expands the capabilities of the AI beyond just text generation. Response fromAI (Discord):** This node sends the AI-generated response back to the Discord channel. It uses the Discord Bot API credentials and replies in a thread under the original message (message_id) in the specified channel_id. Sticky Note1, Sticky Note2, Sticky Note3, Sticky Note4, Sticky Note5, Sticky Note:** These are informational nodes within the workflow providing instructions, code snippets for the Discord bot, and setup guidance for the user. These notes guide the user on setting up the .env file, requirements.txt, the Python bot code, and general recommendations for channel configuration and adding tools. 5️⃣ Setting up Redis Choose a Redis Hosting Provider: You can use a cloud provider like Redis Labs, Aiven, or set up your own Redis instance on a VPS. Obtain Redis Connection Details: Once your Redis instance is set up, you will need the host, port, and password (if applicable). Configure n8n Redis Nodes: In your n8n workflow, configure the "Message History" node with your Redis connection details. Ensure the Redis credential ✅ redis-for-n8n is properly set up with your Redis instance details (host, port, password). 6️⃣ Customizing the Template AI Model:** You can easily swap out the "OpenAI -4o-mini" node with any other AI service supported by n8n (e.g., Cohere, Hugging Face) to use a different language model. Ensure the new language model node is connected to the ai_languageModel input of the "Chat Agent" node. Agent Prompt:** Modify the text parameter in the "Chat Agent" node to change the AI's persona, provide specific instructions, or adjust the response length. Additional Tools:** The "Calculator" node is an example of an AI tool. You can add more Langchain tool nodes (e.g., search, data lookup) and connect them to the ai_tool input of the "Chat Agent" node to extend the AI's capabilities. Refer to the "Sticky Note5" in the workflow for a reminder. Channel Filtering:** Adjust the LISTEN_CHANNELS list in the main.py file of your Discord bot to include or exclude specific Discord channel IDs where the bot should listen for messages. Thread Management:** The "Response fromAI" node can be modified to change how threads are created or managed, or to send responses directly to the channel instead of a thread. The current setup links the response to the original message ID (message_reference). 7️⃣ Testing Instructions Start the Discord Bot: Ensure your main.py script is running on Pella. Activate the n8n Workflow: Make sure your n8n workflow is active and listening for webhooks. Send a Message in Discord: Go to one of the LISTEN_CHANNELS in your Discord server and send a message. Verify Response: The bot should capture the message, send it to n8n, receive an AI-generated response, and post it as a thread under your original message. Check Redis: Verify that the conversation history is being stored and updated correctly in your Redis instance. Look for keys related to user IDs. ✅ Now your bot is running in the background! 🚀

This workflow contains community nodes that are only compatible with the self-hosted version of n8n. This workflow provides a complete, ready-to-use template for a Retrieval-Augmented Generation (RAG) system. It allows you to build a powerful AI chatbot that can answer questions based on the content of PDF documents you provide, using a modern and powerful stack for optimal performance. Good to know Costs:** This workflow uses paid services (OpenAI, Pinecone, Cohere). Costs will be incurred based on your usage. Please review the pricing pages for each service to understand the potential expenses. Video Tutorial (Bahasa Indonesia):** For a step-by-step guide on how this workflow functions, you can watch the accompanying video tutorial here: N8N Tutorial: Membangun Chatbot RAG dengan Pinecone, OpenAI, & Cohere How it works This workflow operates in two distinct stages: 1. Data Ingestion & Indexing: It begins when a .pdf file is uploaded via the n8n Form Trigger. The Default Data Loader node processes the PDF, and the Recursive Character Text Splitter breaks down the content into smaller, manageable chunks. The Embeddings OpenAI node converts these text chunks into vector embeddings (numerical representations). Finally, the Pinecone Vector Store node takes these embeddings and stores (upserts) them into your specified Pinecone index, creating a searchable knowledge base. 2. Conversational AI Agent: A user sends a message through the Chat Trigger. The AI Agent receives the message and uses its VectorDB tool to search the Pinecone index for relevant information. The Reranker Cohere node refines these search results, ensuring only the most relevant context is selected. The user's original question and the refined context are sent to the OpenAI Chat Model (gpt-4.1), which generates a helpful, context-aware answer. The Simple Memory node maintains conversation history, allowing for natural, multi-turn dialogues. How to use Using this workflow is a two-step process: Populate the Knowledge Base: First, you need to add documents. Trigger the workflow by using the Form Trigger and uploading a PDF file. Wait for the execution to complete. You can do this for multiple documents. Start Chatting: Once your data has been ingested, open the Chat Trigger's interface and start asking questions related to the content of your uploaded documents. The Form Trigger is just an example. Feel free to replace it with other triggers, such as a node that watches a Google Drive or Dropbox folder for new files. Requirements To run this workflow, you will need active accounts and API keys for the following services. OpenAI Account & API Key:** Function: Powers text embedding and the final chat generation. Required for the Embeddings OpenAI and OpenAI Chat Model nodes. Pinecone Account & API Key:** Function: Used to store and retrieve your vector knowledge base. Required for the Pinecone Vector Store and VectorDB nodes. You also need to provide your Pinecone Environment. Cohere Account & API Key:** Function: Improves the accuracy of your chatbot by re-ranking search results for relevance. Required for the Reranker Cohere node. Customising this workflow This template is a great starting point. Here are a few ways you can customize it: Change the AI Personality:* Edit the *System Message** in the AI Agent node to change the bot's behavior, tone, or instructions. Use Different Models:** You can easily swap the OpenAI model for another one (e.g., gpt-3.5-turbo for lower costs) in the OpenAI Chat Model node. Adjust Retrieval:** In the VectorDB tool node, you can modify the Top K parameter to retrieve more or fewer document chunks to use as context. Automate Ingestion:** Replace the manual Form Trigger with an automated one, like a node that triggers whenever a new file is added to a specific cloud storage folder.

Who is this for? This template is for any website owner, digital agency, or compliance officer operating within the European Union. It's designed for users who need to comply with the upcoming European Accessibility Act (EAA) but may not have deep technical or legal expertise. Disclaimer This workflow uses an npm package called "cheerio" to work with the specified URLs HTML code. Installing packages is only possible in self hosting. What problem is this workflow solving? / Use Case Starting June 28, 2025, the European Accessibility Act (EAA) mandates that most websites offering products or services in the EU must be accessible and publish a formal Accessibility Statement. Manually creating this legal document is complex, requiring both a technical site analysis and knowledge of specific legal requirements. This workflow automates the generation of a compliant first draft, saving significant time and effort. What this workflow does After you input your details (like website URL and API key) in a central configuration node, this workflow automatically: Scans your live website for accessibility issues using the powerful WAVE API. Processes the scan results to identify the main problem areas. Instructs a Google Gemini AI agent with a specialized legal prompt based on the European Accessibility Act. Generates a formal Accessibility Statement in your desired language. Saves the statement as an .html file and sends it to you as an email attachment. Setup This workflow is designed for a quick setup: Configure All Variables: Click the 'CHANGE THESE: dependencies' node. This is your central control panel. Fill in all the values, including your WAVE API Key, the URL to analyze, company details, and desired output language. Set Up Credentials: You will need to connect your Google accounts for the workflow to run. Gemini: Click the 'gemini 2.5 pro' node, click the gear icon (⚙️) next to the "Credential" field, and connect your Google Gemini API credentials. Gmail: Click the 'Send report by email' node and connect your Gmail account to allow sending the final report. Activate & Execute: Make sure the workflow is active in the top-right corner, then click 'Execute Workflow' to run your first analysis. How to customize this workflow to your needs This template is a great starting point for any EU country. Here's how to adapt it: Localize for Your Country (Important!):* The generated statement contains a placeholder for the "Enforcement Procedure". You *must* edit the prompt in the *'Accessibility Statement Generator'** node to replace this placeholder with the name and link to your specific country's official enforcement body. Change the AI:** Swap the Google Gemini node for any other AI model, like OpenAI or Anthropic Claude, by replacing the node and connecting it to the agent. Change the Trigger:* Replace the *'When clicking ‘Execute workflow’'** node with a Form Trigger or Webhook Trigger to run this workflow based on external inputs, for example, to offer this analysis as a service to your clients.

Who is this for? This workflow is ideal for businesses, accountants, and finance teams who receive bank slip images via LINE and want to automate the extraction of transaction details. It eliminates manual data entry and speeds up financial tracking. What problem does this workflow solve? Many businesses receive bank transfer slips via LINE from customers, but manually recording transaction details into spreadsheets is time-consuming and error-prone. This workflow automates the entire process, extracting structured data from the bank slips and storing it in Google Sheets for seamless record-keeping. What this workflow does: Receives bank slip images from LINE BOT Extracts transaction details (sender, receiver, amount, transaction ID) using SpaceOCR Automatically logs extracted data into Google Sheets Works with Standard Bank Slips & PromptPay transactions Eliminates manual data entry and reduces errors Setup Instructions: 1. Prerequisites A LINE BOT with Messaging API enabled A SpaceOCR API Key (Get from https://spaceocr.com/) A Google Sheets account to store extracted data An n8n instance running (Cloud or Self-hosted) 2. Setup Google Sheets Create a Google Sheet with the following structure: A (Date) B (Time) C (Sender) D (Receiver) E (Bank Name) F (Amount) G (Transaction ID) Ensure your Google Sheets API is enabled and connected to n8n. For an example of the required format, check this Google Sheets template: Google Sheets Template 3. Configure n8n Workflow 1. Webhook Node (Receives bank slip from LINE BOT) Set method:* Set Path:* 2. HTTP Request (Download Image from LINE Message) Retrieves image URL from the LINE message payload 3. SpaceOCR Node (Extract Text from Bank Slip) Input:* API Key:* #### 4. Google Sheets Node (Save Transaction Data) Select your Google Sheet Map extracted data (sender, receiver, amount, etc.) to the respective columns 4. Deploy & Test Activate the workflow in n8n Set Webhook URL in LINE Developer Console Send a test bank slip image to the LINE BOT Check Google Sheets for extracted transaction data

Generate and Publish AI Content to LinkedIn and X (Twitter) with n8n Overview This n8n workflow automates the generation and publishing of AI-powered social media content across LinkedIn and X (formerly Twitter). By leveraging AI, this workflow helps social media managers, marketers, and content creators streamline their posting process. Who is this for? Social media managers Content creators Digital marketers Businesses looking to automate content generation Features AI-powered content creation** tailored for LinkedIn and X (Twitter) Automated publishing** to both platforms Structured output parsing** to ensure consistency OAuth2 authentication** for secure posting Merge and confirmation steps** to track successful postings Setup Instructions Prerequisites Before using this workflow, ensure you have: An n8n instance set up API credentials for: Google Gemini AI (for content generation) X Developer Account with OAuth2 authentication LinkedIn Developer Account with OAuth2 authentication A form submission service integrated with n8n Workflow Breakdown 1. Trigger: Form Submission A user submits a form containing the post title. The form is secured with Basic Authentication. The submitted title is passed to the AI Agent. 2. AI Content Generation The Google Gemini Chat Model processes the title and generates: LinkedIn post content Twitter (X) post content Hashtags Call-to-action (LinkedIn) Character limit check (Twitter) 3. Parsing AI Output A structured output parser converts the AI-generated content into a JSON format. Ensures correct formatting for LinkedIn and Twitter (X). 4. Publishing to Social Media X (Twitter) Posting Extracts the Twitter post from the AI output. Publishes it via an OAuth2-authenticated X (Twitter) account. LinkedIn Posting Extracts the LinkedIn post from the AI output. Publishes it via an OAuth2-authenticated LinkedIn account. 5. Merging Post Results Merges the response data from both LinkedIn and Twitter after publishing. 6. Confirmation Step Displays a final confirmation form once the posts are successfully published. Benefits Save time** by automating content creation and publishing. Ensure consistency** across platforms with structured AI-generated posts. Secure authentication** using OAuth2 for LinkedIn and Twitter. Increase engagement** with AI-optimized hashtags and CTAs. This workflow enables seamless social media automation, helping professionals post engaging AI-powered content effortlessly. 🚀

Built this for a dedicated Slack outage-notifications channel — works well on both desktop and mobile. This is for: IT Administrators & small MSPs looking to streamline M365 alerts from one or multiple mailboxes into a single or specific Slack channels IT Admins who prefer ChatOps over management-by-email What does it do Scans for M365 outage alerts emails (every 1 min) Checks if it impacts a specific user region (if the alert calls it out, countries have to be manually set) Summarizes the incident using OpenAI o4-mini (cheap model - or you can swap for local Ollama) Sends a Slack Block to your outage channel with incident link (can be extended) Deletes the original alert email after successful delivery Credentials Outlook: Create an Outlook credential (OAuth2.0) to point to the mailbox (regular or shared) where M365 service alerts will be received Slack: Create a Slack bot credential with access to the slack channel you want updates posted to OpenAI: Create a OpenAI credential that has access to the GPT-4O-MINI model. Recommend you use projects in OpenAI so that you may set a per-project-budget and not impact other projects. Review this OpenAI documentation for more info on managing Projects in the API portal. Expect this to consume no more than 1-2 cents per month on average. Setup Download & import the workflow Modify the first Outlook block (Check for 365 Service Alert) to use the Outlook credential Modify the OpenAI block's system prompt to call out the countries your users reside in ie. "- Assume the organization has users primarily in the U.S. and Australia. If those regions are affected, state: "Your users may have been affected." Otherwise, add: "No impact expected for your user base."" ← swap U.S. & Australia for desired countries Modify the Slack block (Post outage to Slack) to specify the channel updates will be posted to Sample Slack Output Workflow Diagram

🤖 AI Customer Support Agent with Google Docs Knowledge (Telegram + OpenAI) This no-code workflow turns your Telegram bot into an intelligent, always-on AI support agent that references your business documentation in Google Docs to respond to customer queries—instantly and accurately. Watch full step-by-step video tutorial of the build here: https://youtu.be/Mlv7CjGO7wI 🔧 How it works: Telegram Trigger – Captures incoming messages from users on your Telegram bot Langchain AI Agent (OpenAI GPT) – Interprets the message and uses RAG (retrieval-augmented generation) techniques to craft an answer Google Docs Tool – Connects to and retrieves context from your specified Google Doc (e.g. FAQ, SOPs, policies) Memory Buffer – Keeps track of recent chat history for more human-like conversations Telegram Reply Node – Sends the AI-generated response back to the user 💡 Use Cases: E-commerce customer service SaaS product onboarding Internal helpdesk bot for teams WhatsApp-style support for digital businesses 🧠 What makes this powerful: Supports complex questions by referencing a live Google Doc knowledge base Works in plain conversational language (no buttons or forms needed) Runs 24/7 with zero code Easily extendable to Slack, WhatsApp, or email support 🛠️ Tools used: Telegram Node (trigger + send) Langchain Agent with OpenAI GPT Google Docs Tool Memory Buffer Sticky Notes for easy understanding

Overview Slack quietly released an update to their API that allows developers to build "AI Apps & Agents", which is a special classification of apps that have access to several special capabilities including: Multiple simultaneous chat threads with one user Loading "three dots" UI while your agent is thinking Option for users to pin your app to their top bar for quick chat access This workflow demonstrates how to build a Slack agent that takes advantage of all of these features. For a full video walkthrough of this workflow, watch this YouTube tutorial. Setup Instructions All of the below steps are required for this workflow to function properly unless otherwise noted. Create a Slack App Visit api.slack.com and click "Your Apps" Create a new app from scratch and follow the setup instructions In the Agents & AI Apps tab, enable the toggle and give your app a brief description In the OAuth & Permissions tab, enable the following bot token scopes: assistant:write chat:write channels:read im:history Install the app into your workspace and grant the requested permissions In your Slack workspace, right click your app's name in the sidebar, click "View app details", and make note of your apps Channel ID - you'll need this later. Copy your app's Bot User OAuth Token - you'll need that to create your n8n credentials In the Event Subscriptions tab, enable events and paste the workflows PRODUCTION webhook url (from this workflow's trigger node) into the input. In the same tab under "Susbcribe to bot events", select message.im Create a Postgres database In order to save the chat history and give your agent a working memory, you'll need your own Postgres database. You can use Supabase, Neon, or any other Postgres database provider. Once you've added your database's credentials to n8n, you can select those credentials in the Postgres Chat Memory node. This worklow saves all chat history in a table called chat_histories, but you name the table whatever you want. Create n8n Credentials You'll need to create the following credentials: Slack API. Use your Bot User OAuth Token referenced above. Bearer Auth. Use the same Bot User OAuth Token. Postgres. Use the connection string or config from your database provider. OpenRouter (or any other LLM model for the agent's model node) Wire Everything Up Now that you've created your Slack app, have your Postgres database, and have created credentials, follow these steps to wire up your workflow: In the "On Message Received" trigger, use your Slack API credential and enter your apps Channel ID in the "Channel To Watch" field. In the "Set Thinking Status" node, use your Bearer Auth credential. In the "Postgres Chat Memory" node, use your Postgres credential. In the "Send Reply" node, use your Slack API credential. Using the Chatbot Once you've completed the setup process and added in your credentials, you'll have a fully functional Slack chatbot complete with threads, loading UI, and the ability to pin your app to your workspace's top bar. Taking the Next Steps Now that this skeleton app is in place, it's up to you to add horsepower to the AI agent at the center of it all. Customize the prompts and add whatever tools you'd like. The sky is the limit! If you have any questions or feedback about this workflow, or would like me to build custom workflows for your business, email me at n8n@paperjam.agency.

Who is this template for? This workflow template is designed for people seeking alerts when certain specific changes are made to any web page. Leveraging agentic AI, it analyzes the page every day and autonomously decides whether to send you an e-mail notification. Example use cases Track price changes on [competitor's website]. Notify me when the price drops below €50. Monitor new blog posts on [industry leader's website] and summarize key insights. Check [competitor's job page] for new job postings related to software development. Watch for new product launches on [e-commerce site] and send me a summary. Detect any changes in the terms and conditions of [specific website]. Track customer reviews for [specific product] on [review site] and extract key themes. How it works When clicking 'test workflow' in the editor, a new browser tab will open where you can fill in the details of your espionage assignment Make sure you be as concise as possible when instructing AI. Instruct specific and to the point (see examples at the bottom). After submission, the flow will start off by extracting both the relevant website url and an optimized prompt. OpenAI's structured outputs is utilized, followed by a code node to parse the results for further use. From here on, the endless loop of daily checks will begin: Initial scrape 1 day delay Second scrape AI agent decides whether or not to notify you Back to step 1 You can cancel an espionage assignment at any time in the executions tab Set up steps Insert your OpenAI API key in the structured outputs node (second one) Create a Firecrawl account and connect your Firecrawl API key in both 'Scrape page'-nodes Connect your OpenAI account in the AI agents' model node Connect your Gmail account in the AI agents' Gmail tool node

✉️ Telegram Email Agent with GPT + Gmail Category: Messaging / AI Agent Level: Beginner-Friendly Tags: Telegram, Email Automation, AI Agent, Gmail, GPT Model Watch Step-by-step video guide here: https://www.youtube.com/watch?v=nyI40s9QOuw&t=420s&pp=0gcJCb4JAYcqIYzv 🤖 What This Workflow Does This workflow turns your Telegram bot into a personal email assistant powered by AI. With just a message on Telegram, users can: Send an email via Gmail Automatically generate the email content using OpenAI Models. Get confirmation or responses directly in Telegram It's like ChatGPT meets Gmail, inside your Telegram chat. 🔧 How It Works Telegram Trigger – Listens for incoming messages from your bot. AI Agent – Processes the input using OpenAI Model and converts it into structured email content (To, Subject, Body). Memory Node – Stores short-term context per user (via chat ID), so the agent can hold simple conversations. Gmail Node – Sends the generated email using your Gmail account. Telegram Node – Replies to the user confirming the output or status. 🧠 Why This is Useful Ever wanted to send an email while on the go, without typing the whole thing out in Gmail? This is a fast, intuitive, and AI-powered way to: Dictate or draft emails from anywhere Create an AI-powered virtual assistant via Telegram Integrate n8n's Langchain Agent with real-world productivity use cases 🪜 Setup Instructions Connect your Telegram bot via BotFather and add the credentials in n8n. Set up your OpenAI API key (GPT-4o-mini recommended). Add your Gmail OAuth credentials. Activate the workflow and start messaging your bot!

Use case This automation is for teams working in Notion. When you have a lot of back and forth in the comment section, it’s easy to lose track of what is going on in the conversation. This automation relies on AI to generate a summary of the comment section. How it works Every hour (the trigger can be adapted to your need and usecase), the automation checks if new comments have been added to the pages of your Notion database. If there are new comments, the comments are sent to an AI model to write a summary. The summary is then added to a predefined page property. The automation also updates a “Last execution” property. This prevents to re-generate the AI summary when no new comments have been received. Setup Define your Notion variables: Notion database, property that will hold the AI summary, property that will hold the last execution date of the automation. Set up your Notion credentials. Set up your AI model credentials (API key). How to adjust it to your needs Use the LLM model of your choice. In this template, I used Gemini but you can easily replace it by ChatGPT, Claude, etc. Adapt the prompt to your use case to get better summaries: specify the maximum number of characters, give an example, etc. Adapt the trigger to your needs. You could use Notion webhooks as trigger in order to run the automation only when a new comment is added (this setup is advised if you’re on n8n cloud version).

🎯 What It Does: This project lets you talk to Siri (via Apple Shortcuts) and record or query your daily spending. The shortcut sends your message to an n8n Webhook, which uses AI to decide whether it’s for writing or reading finance data, then replies with a human-friendly message — all powered by n8n + AI + Google Sheets. ⸻ 🌐 PART 1: n8n Setup 🧩 1. Create a Webhook Trigger in n8n • Add a node: Webhook • Set HTTP Method: POST • Set Path: siri-finance • Enable “Respond to Webhook” = ✅ 🧠 2. Add AI Agent Node (e.g. OpenAI, Ollama, Gemini) • Use system prompt like: You are a finance assistant. Decide if the user wants to record or read transactions. If it's recording, return a JSON object with date, type, name, amount, and expense/income. If it's reading, return date range and type (Expense/Income). Always reply with a human-friendly summary. • Input: {{ $json.text }} (from webhook) • Output: structured json.output 🧮 3. (Optional) Add Logic to write to DB / Supabase / Google Sheets • Append tool: Adds a new row • Read tool: Queries past data Now your n8n flow is ready! ⸻ 📱 PART 2: iOS Shortcut Setup ⚙️ 1. Create a new Shortcut • Name it: 記帳助理 (or Finance Bot) • Add Action: Ask for Input • Prompt: “請說出你的記帳內容” • Input Type: Text • Add Action: Get Contents of URL • Method: POST • URL: https://your-n8n-domain/webhook/siri-finance • Headers: Content-Type: application/json • Request Body: { "text": "Provided Input" } • Replace "Provided Input" with Magic Variable → Input Result 🔊 2. Show Result • Add Action: Show Result • Content: Get Contents of URL 🗣️ 3. Optional: Add “Speak Text” • If you want Siri to speak it back, add Speak Text after Show Result. ⸻ ✅ Example Usage • You: “Hey Siri, 開支$50 早餐” • Siri: “已記錄支出：項目 早餐，金額 $50，已寫入” Or • You: “查一下我過去7日用了幾多錢” • Siri: “你過去7日總支出為 $7684.64，包括：⋯⋯” ⸻ 📦 Files to Share You can package the following: • .shortcut file export • Sample n8n workflow .json • Optional Supabase schema / Google Sheet template ⸻ 💡 Tips for Newcomers • Keep your Webhook public but protect with token if needed. • Ensure you handle emoji and newline safely for iOS compatibility. • Add logging nodes in n8n to help debug Siri messages. ⸻ 🗣️ Optional Project Name “Siri 記帳助理” / “Finance VoiceBot” A simple voice-first way to manage your daily expenses.