Theo Marcadet
Just like every business has its own processes, every business will have its own AI agents — I sell and teach that transition From France 🇫🇷 — Discipline, Humility, Benevolence
Templates by Theo Marcadet
Autonomous customizable support chatbot on Intercom + Discord thread reports
Click here to access this Workflow for free. --- Connect your own LLM-boosted chatbot to Intercom (f* their overly priced FlN Agent), and stay in touch on Discord This workflow connects your Intercom chat system with your own AI Agent and sends a complete log of each conversation to Discord using threads. It allows you to run a fully automated support system while maintaining full visibility of the bot's behavior in real time. For every new conversation in Intercom, a thread is created in a specified Discord channel. Each message from the user, the AI, and even manual human responses is logged to the thread, offering full traceability and transparency. You also have fine-grained control over the AI agent. By simply clicking the ⭐️ star in Intercom’s UI, support agents can instantly pause or resume AI responses for a specific chat — no coding or config changes needed. --- Requirements A working n8n instance An Intercom account with permission to set up webhooks A Discord bot with the following permissions: Send Messages Create Public/Private Threads Manage Threads API credentials for your preferred LLM (OpenAI is used by default) Google Chrome or any browser to access Intercom’s UI --- Setup Intercom: Go to Intercom’s webhook settings. Add a webhook that listens to new incoming messages and points to the Webhook URL in this n8n workflow. Make sure to send full conversation data. Discord: Create a Discord bot and invite it to your server with the required permissions. In the Discord + Token node at the top of the workflow: Add your bot token Add the ID of the channel where threads should be created LLM / AI Agent: By default, the workflow uses OpenAI via HTTP Request. You can substitute it with any LLM provider of your choice. Make sure to set up your credentials in n8n and select them in the HTTP nodes. HTTP Authentication Tips: For both Intercom and Discord API calls, it's recommended to create credentials in n8n's Credential Manager. Then, assign those credentials inside each HTTP Request node for a cleaner setup. --- Usage When a new conversation starts in Intercom, a Discord thread is created automatically. Each message — user input, AI response, and human reply — is logged into the Discord thread in real time. The AI replies automatically unless the ⭐️ star is checked in Intercom: ☆ Unchecked = AI replies enabled ⭐️ Checked = AI replies disabled, human takeover enabled This gives you on-the-fly control of each conversation’s automation level directly from the Intercom inbox. --- Customization You can replace OpenAI with any LLM that provides a compatible API. Discord channel ID, thread naming, and message formatting can be customized to match your team’s preferences. You can expand the workflow to handle events like conversation closure or satisfaction ratings for deeper analytics. --- If you need any help, or have any question, feel free to come discuss about it on Telegram
By Theo MarcadetAutomated n8n workflow backup system with Google Drive and archiving
Click here to access this Workflow for free. --- Automatically save and archive your workflows This workflow provides a fully automated system for backing up your n8n workflows to Google Drive. It runs on a schedule (which you can customize) and saves each workflow as a .json file, named after the workflow itself. All active workflows are backed up by default and stored in a specified Google Drive folder of your choice. A special feature of this workflow is its ability to archive and clean up old workflows. If you tag a workflow with ARCHIVE (case-sensitive), the workflow will be backed up into a designated Archive subfolder inside your main Google Drive folder — and then automatically deleted from your n8n instance. This keeps your workspace clean and reduces clutter while ensuring everything is safely stored. Requirements A Google Drive account connected to n8n (via OAuth2 or preferred method) An n8n instance with access to your workflows A folder in Google Drive for storing backups Inside this folder, create a subfolder named Archive Folder IDs for both the main and archive folders (Find it in folder's URL e.g. https://drive.google.com/drive/u/0/folders/1F7Exr2xkZQFvmHmjdFD43K0jfd) Setup In your Google Drive, create a main folder for storing backups. Inside it, create a subfolder called Archive. Retrieve the folder IDs for both folders In n8n, connect your Google Drive credentials. Open the workflow and insert the folder IDs in the appropriate nodes. Optionally, edit the schedule trigger to set your preferred backup frequency. Usage Once configured, the workflow will: Run automatically on the defined schedule Back up all existing workflows into your Drive For workflows tagged with ARCHIVE: Save them into the Archive subfolder Delete them from your n8n instance The workflow ensures your automation logic is safely versioned, with optional cleanup for outdated or unused flows. Customization You can modify the schedule to run daily, weekly, or at any custom interval using the built-in Cron node. You can also change the folder structure or naming conventions in the Google Drive nodes to match your personal or team preferences. Workflows are archived only if they are explicitly tagged with ARCHIVE — this tag is case-sensitive. --- If you need any help, or have any question, feel free to come discuss about it on Telegram
By Theo MarcadetExport Wordpress to PineCone vector store
Click here to access this Workflow for free. --- Make your website the knowledge base of your LLM chatbot This workflow automatically syncs your WordPress content (Pages and Posts) into a vector database like Pinecone, making it searchable and usable by AI agents through embeddings. This allows your AI chatbot to stay continuously up to date with your latest site content — without manual uploads or retraining steps. It pulls all published and private content via WordPress's REST API, processes the data into clean JSON, and sends it to Pinecone using vector embeddings (OpenAI by default, but swappable). This enables a dynamic AI knowledge base that grows with your website. Best of all, you don’t need to modify your workflow: keep managing your content in WordPress — this automation takes care of the rest. --- Requirements A WordPress site with API access Pinecone account credentials (API Key + Environment) OpenAI account (or any LLM with embedding capabilities) An n8n instance (cloud or self-hosted) --- Setup Configure WordPress Access: Create HTTP credentials in n8n with your WordPress site's base URL and authentication. This workflow uses the REST API to fetch pages and posts. Connect Pinecone and OpenAI: Set up credentials for both Pinecone and OpenAI in n8n's Credentials Manager. The workflow automatically uses them in the relevant nodes. Set Your Schedule: The workflow starts with a Schedule node. Open it to adjust how often the sync runs (e.g., daily, weekly). Check the Nodes: Make sure your folder structure and API URLs match your WordPress setup. Adjust filters if needed (only published and private content is fetched by default). --- Usage Once the workflow is active: It runs on a defined schedule. It fetches all WordPress Pages and Posts. It converts them into structured JSON, including metadata (title, tags, URL, categories, etc.). It generates embeddings (via OpenAI) and sends them to Pinecone. All vectors are stored with rich metadata for easy use in search/chat-based AI tools. You’ll have a continuously refreshed, vectorized copy of your WordPress site content. --- Customization Replace Pinecone: You can swap Pinecone for any other vector database by updating the vector node. Swap Embedding Provider: You’re not limited to OpenAI — just replace the embedding step with your LLM of choice. Selective Sync: If you want to sync only posts or only pages: Delete the node that fetches the type you don’t want. Remove the associated Merge node. Schedule: Edit the first node to control how often the automation runs. --- If you need any help, or have any question, feel free to come discuss about it on Telegram
By Theo Marcadet