Basic Automatic Gmail Email Labelling with OpenAI and Gmail API

Automatically categorize incoming Gmail emails using the power of OpenAI's language models. This workflow connects your Gmail account to OpenAI, allowing you to intelligently label emails based on their content and your predefined criteria. It starts with a Gmail Trigger that monitors for new emails, then uses an OpenAI Chat Model to analyze the email's subject and body. A sophisticated Gmail labelling agent, backed by an OpenAI agent and a Window Buffer Memory for context, determines the most appropriate label. If a label doesn't exist, the workflow can even create it via the Gmail - create label node before applying it to the message using Gmail - add label to message. This is ideal for marketing teams needing to sort customer inquiries, support departments organizing tickets, or anyone overwhelmed by their inbox who wants to automate email management and reduce manual sorting time. By automating this process, you save significant time and effort previously spent on manually sifting through and categorizing emails, ensuring your inbox remains organized and important messages are easily discoverable.
13 nodesmanual trigger98 views0 copiesMarketing
GmailOpenAI

Workflow JSON

{"nodes": [{"id": "2a41e2da-19f7-4c31-ab93-3a534db3179e", "name": "Gmail Trigger", "type": "n8n-nodes-base.gmailTrigger", "position": [-360, -260], "parameters": {"filters": {}, "pollTimes": {"item": [{"mode": "everyX", "unit": "minutes", "value": 5}]}}, "credentials": {"gmailOAuth2": {"id": "", "name": "[Your gmailOAuth2]"}}, "typeVersion": 1.2}, {"id": "a25e0e42-8eab-49c5-a553-797da40eb623", "name": "OpenAI Chat Model1", "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi", "position": [-220, -60], "parameters": {"options": {"maxTokens": 4096}}, "credentials": {"openAiApi": {"id": "", "name": "[Your openAiApi]"}}, "notesInFlow": false, "typeVersion": 1}, {"id": "cf437748-a0df-42a2-b1ca-f93162d85bfe", "name": "Gmail - read labels", "type": "n8n-nodes-base.gmailTool", "position": [80, -40], "webhookId": "d8ec9401-a9ff-4fe2-9c1e-5a8036cd96c9", "parameters": {"resource": "label", "returnAll": true, "descriptionType": "manual", "toolDescription": "Tool to read all existing gmail labels"}, "credentials": {"gmailOAuth2": {"id": "", "name": "[Your gmailOAuth2]"}}, "typeVersion": 2.1}, {"id": "152f1970-7a1f-4977-9c21-64b69242d3a9", "name": "Gmail - get message", "type": "n8n-nodes-base.gmailTool", "position": [260, -40], "webhookId": "d8ec9401-a9ff-4fe2-9c1e-5a8036cd96c9", "parameters": {"messageId": "={{ $fromAI('gmail_message_id', 'id of the gmail message, like 1944fdc33f544369', 'string') }}", "operation": "get", "descriptionType": "manual", "toolDescription": "Tool to read a specific message based on the message ID"}, "credentials": {"gmailOAuth2": {"id": "", "name": "[Your gmailOAuth2]"}}, "typeVersion": 2.1}, {"id": "ae09cedc-9675-4080-bcdc-3d6c4e4bc490", "name": "Gmail - add label to message", "type": "n8n-nodes-base.gmailTool", "position": [460, -40], "webhookId": "7a87b026-1c6e-40e1-a062-aefdd1af1585", "parameters": {"labelIds": "={{ $fromAI('gmail_categories', 'array of label ids') }}", "messageId": "={{ $fromAI('gmail_message_id') }}", "operation": "addLabels", "descriptionType": "manual", "toolDescription": "Tool to add label to message"}, "credentials": {"gmailOAuth2": {"id": "", "name": "[Your gmailOAuth2]"}}, "typeVersion": 2.1}, {"id": "be4a92ab-d3ab-451b-8655-172851f68628", "name": "Gmail - create label", "type": "n8n-nodes-base.gmailTool", "position": [640, -40], "webhookId": "d8ec9401-a9ff-4fe2-9c1e-5a8036cd96c9", "parameters": {"name": "={{ $fromAI('new_label_name', 'new label name', 'string' ) }} ", "options": {}, "resource": "label", "operation": "create", "descriptionType": "manual", "toolDescription": "Tool to create a new label, only use if label does not already exist"}, "credentials": {"gmailOAuth2": {"id": "", "name": "[Your gmailOAuth2]"}}, "typeVersion": 2.1}, {"id": "a40466d2-2fe3-4a97-98fe-b14cc38cc141", "name": "Gmail labelling agent", "type": "@n8n/n8n-nodes-langchain.agent", "notes": "Objective:\nAutomatically categorize incoming emails based on existing Gmail labels or create a new label if none match.\n\nTools:\n- Get message\n- Read all labels\n- Create label\n- Assign label to message\n\nInstructions:\n\nLabel Matching:\n\nAnalyze the email's subject, sender, recipient, keywords, and content.\nCompare with existing Gmail labels to find the most relevant match.\nLabel Assignment:\n\nAssign the email to the most appropriate existing label.`\nRemove the inbox label if the email is of less importance (like ads, promotions, aka \"Reclame\"), keep normal and important emails in the inbox.\nIf no suitable label exists, create a new label based on the existing labels. Try reusing existing labels as much as possible. Always create a label as a sublabel, if no label applies, if the main label already exists, create the new label under the existing label, if no main label exists, create the label AI and create the new label under this label.\nLabel Creation:\n\nEnsure new labels align with the structure of existing ones, including capitalization, delimiters, and prefixes.\nExamples:\n\nIf the email subject is \"Project Alpha Update,\" assign to [Project Alpha] if it exists.\nFor \"New Vendor Inquiry,\" create \"Vendor Inquiry\" if no relevant label exists.\nOutcome:\nEmails are consistently categorized under the appropriate or newly created labels, maintaining Gmail's organizational structure.", "onError": "continueErrorOutput", "position": [-60, -260], "parameters": {"text": "=Label the email based on the details below:\n{{ JSON.stringify($json) }}", "options": {"maxIterations": 5, "systemMessage": "Objective:\nAutomatically categorize incoming emails based on existing Gmail labels or create a new label if none match.\n\nTools:\n- Get message\n- Read all labels\n- Create label\n- Assign label to message\n\nInstructions:\n\nLabel Matching:\n\nAnalyze the email's subject, sender, recipient, keywords, and content.\nCompare with existing Gmail labels to find the most relevant match.\nLabel Assignment:\n\nAssign the email to the most appropriate existing label.`\nRemove the inbox label if the email is of less importance (like ads, promotions, aka \"Reclame\"), keep normal and important emails in the inbox.\nIf no suitable label exists, create a new label based on the existing labels. Try reusing existing labels as much as possible. Always create a label as a sublabel, if no label applies, if the main label already exists, create the new label under the existing label, if no main label exists, create the label AI and create the new label under this label.\nLabel Creation:\n\nEnsure new labels align with the structure of existing ones, including capitalization, delimiters, and prefixes.\nExamples:\n\nIf the email subject is \"Project Alpha Update,\" assign to [Project Alpha] if it exists.\nFor \"New Vendor Inquiry,\" create \"Vendor Inquiry\" if no relevant label exists.\nOutcome:\nEmails are consistently categorized under the appropriate or newly created labels, maintaining Gmail's organizational structure."}, "promptType": "define"}, "notesInFlow": true, "retryOnFail": false, "typeVersion": 1.7}, {"id": "6b514df4-761c-4072-abf8-d572ee4b8030", "name": "Window Buffer Memory", "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow", "position": [-60, -40], "parameters": {"sessionKey": "={{ $json.id }}", "sessionIdType": "customKey"}, "typeVersion": 1.3}, {"id": "f06717ed-00d7-4a99-a78c-53217a0067e7", "name": "Wait", "type": "n8n-nodes-base.wait", "position": [-220, -260], "webhookId": "2066b863-4526-40cf-90aa-82229895a73c", "parameters": {"amount": 1}, "typeVersion": 1.1}, {"id": "f6084fc3-2b6b-488f-b212-f179435e1a63", "name": "Sticky Note", "type": "n8n-nodes-base.stickyNote", "position": [-640, -300], "parameters": {"content": "## Gmail trigger\nPoll Gmail every x minutes, trigger when a new email is received.\n\n- Gmail API"}, "typeVersion": 1}, {"id": "5ede55a4-52ae-48c0-969e-afa45d19f2f0", "name": "Sticky Note1", "type": "n8n-nodes-base.stickyNote", "position": [380, -960], "parameters": {"width": 780, "height": 840, "content": "## Gmail labelling agent\n- Read the message\n- Read existing labels\n- Create a new label if needed\n- Assign label to message\n\n----\n\nObjective:\nAutomatically categorize incoming emails based on existing Gmail labels or create a new label if none match.\n\nTools:\n- Get message\n- Read all labels\n- Create label\n- Assign label to message\n\nInstructions:\n\nLabel Matching:\n\nAnalyze the email's subject, sender, recipient, keywords, and content.\nCompare with existing Gmail labels to find the most relevant match.\nLabel Assignment:\n\nAssign the email to the most appropriate existing label.`\nRemove the inbox label if the email is of less importance (like ads, promotions, aka \"Reclame\"), keep normal and important emails in the inbox.\nIf no suitable label exists, create a new label based on the existing labels. Try reusing existing labels as much as possible. Always create a label as a sublabel, if no label applies, if the main label already exists, create the new label under the existing label, if no main label exists, create the label AI and create the new label under this label.\nLabel Creation:\n\nEnsure new labels align with the structure of existing ones, including capitalization, delimiters, and prefixes.\nExamples:\n\nIf the email subject is \"Project Alpha Update,\" assign to [Project Alpha] if it exists.\nFor \"New Vendor Inquiry,\" create \"Vendor Inquiry\" if no relevant label exists.\nOutcome:\nEmails are consistently categorized under the appropriate or newly created labels, maintaining Gmail's organizational structure."}, "typeVersion": 1}, {"id": "7c8bb6de-b729-4c8e-90c2-641d173ed3dd", "name": "Sticky Note2", "type": "n8n-nodes-base.stickyNote", "position": [160, 160], "parameters": {"width": 440, "content": "## Gmail API\n- Add credentials "}, "typeVersion": 1}, {"id": "e9d05013-9546-426f-bdc7-45199dbfc72a", "name": "Sticky Note3", "type": "n8n-nodes-base.stickyNote", "position": [-580, 80], "parameters": {"width": 440, "content": "## OpenAI\n- Add credentials "}, "typeVersion": 1}], "pinData": {}, "connections": {"Wait": {"main": [[{"node": "Gmail labelling agent", "type": "main", "index": 0}]]}, "Gmail Trigger": {"main": [[{"node": "Wait", "type": "main", "index": 0}]]}, "OpenAI Chat Model1": {"ai_languageModel": [[{"node": "Gmail labelling agent", "type": "ai_languageModel", "index": 0}]]}, "Gmail - get message": {"ai_tool": [[{"node": "Gmail labelling agent", "type": "ai_tool", "index": 0}]]}, "Gmail - read labels": {"ai_tool": [[{"node": "Gmail labelling agent", "type": "ai_tool", "index": 0}]]}, "Gmail - create label": {"ai_tool": [[{"node": "Gmail labelling agent", "type": "ai_tool", "index": 0}]]}, "Window Buffer Memory": {"ai_memory": [[{"node": "Gmail labelling agent", "type": "ai_memory", "index": 0}]]}, "Gmail - add label to message": {"ai_tool": [[{"node": "Gmail labelling agent", "type": "ai_tool", "index": 0}]]}}}

How to Import This Workflow

  1. 1Copy the workflow JSON above using the Copy Workflow JSON button.
  2. 2Open your n8n instance and go to Workflows.
  3. 3Click Import from JSON and paste the copied workflow.

Don't have an n8n instance? Start your free trial at n8nautomation.cloud

Related Templates

Automate Blog Creation in Brand Voice with AI

Generate blog posts in your brand's unique voice and style directly within WordPress using this powerful AI-driven workflow. It begins by fetching existing articles from your blog via an HTTP request, then intelligently extracts their content and structure using an HTML node and an AI chain LLM to understand your established writing patterns. Concurrently, an AI information extractor analyzes these articles to pinpoint and define your brand's distinct voice characteristics. This extracted style and voice, along with the structural insights, are then fed into an AI content generation agent which crafts new blog post drafts. Finally, these AI-generated articles are automatically saved as drafts in your WordPress instance, ready for review and publication. This workflow is ideal for marketing teams, content creators, and agencies looking to scale their content production while maintaining consistent brand messaging and reducing the manual effort involved in drafting new posts, ultimately saving significant time and resources.

27 nodes

Auto-Tag Blog Posts in WordPress with AI

Automatically tag your WordPress blog posts with relevant keywords using the power of AI. This n8n workflow connects your WordPress site with OpenAI's advanced language models to intelligently analyze new or existing articles and assign appropriate tags, streamlining your content management process. It starts by either triggering manually or by an RSS Feed Trigger monitoring your blog for new content, then sends the article text to an OpenAI Chat Model for tag generation. The workflow then checks your existing WordPress tags via an HTTP Request to avoid duplicates, creates any new tags needed with another HTTP Request, and finally updates the WordPress post with the newly generated and existing tags. This automation is perfect for content managers, marketing teams, and bloggers who want to improve SEO, enhance content discoverability, and reduce the manual effort of categorizing a large volume of articles, ultimately saving significant time and ensuring consistent tagging across your entire blog.

32 nodes

Extract spend details (template)

Automate the extraction and tracking of spend details directly from your email inbox into Google Sheets. This workflow connects Gmail to automatically retrieve invoice and payment emails, then uses AI models like Google Gemini Chat Model and Groq Chat Model to intelligently parse and extract key financial data. The extracted information, including vendor, amount, date, and itemized details, is then seamlessly organized and sent to Google Sheets for centralized record-keeping. This powerful automation is ideal for marketing teams, small businesses, and freelancers who need to efficiently monitor expenses, reconcile accounts, and gain better financial visibility without manual data entry, saving significant time and reducing errors in financial tracking.

24 nodes

Ready to automate with n8n?

Get affordable managed n8n hosting with 24/7 support.