Automate Email Responses with Gmail API and Anthropic Claude AI

{
  "id": "uf6t5qhnQMgilteE",
  "meta": {
    "instanceId": "[REDACTED_INSTANCE_ID]",
    "templateCredsSetupCompleted": true
  },
  "name": "Email Manager",
  "tags": [
    {
      "id": "[REDACTED_TAG_ID]",
      "name": "Personal Assistant",
      "createdAt": "2025-05-28T13:58:43.876Z",
      "updatedAt": "2025-05-28T13:58:43.876Z"
    }
  ],
  "nodes": [
    {
      "id": "5e01a927-caf7-41ad-aca7-d7504e313564",
      "name": "Gmail Trigger",
      "type": "n8n-nodes-base.gmailTrigger",
      "position": [
        -20,
        -60
      ],
      "parameters": {
        "filters": {},
        "pollTimes": {
          "item": [
            {
              "mode": "everyMinute"
            }
          ]
        }
      },
      "credentials": {
        "gmailOAuth2": {
          "id": "[REDACTED_GMAIL_CRED_ID]",
          "name": "Gmail account"
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "a8644856-a8bf-49de-84a7-3fd9f93550d3",
      "name": "Gmail",
      "type": "n8n-nodes-base.gmail",
      "position": [
        200,
        -20
      ],
      "webhookId": "[REDACTED_WEBHOOK_ID]",
      "parameters": {
        "simple": false,
        "options": {},
        "messageId": "={{ $json.id }}",
        "operation": "get"
      },
      "credentials": {
        "gmailOAuth2": {
          "id": "[REDACTED_GMAIL_CRED_ID]",
          "name": "Gmail account"
        }
      },
      "typeVersion": 2.1
    },
    {
      "id": "f6ddae9f-303b-46cd-980d-1d46fc4ecbec",
      "name": "Structured Output Parser",
      "type": "@n8n/n8n-nodes-langchain.outputParserStructured",
      "position": [
        1000,
        100
      ],
      "parameters": {
        "jsonSchemaExample": "{\n\t\"label\": \"Label name\", \n\"label ID\": \"label ID\" \n}"
      },
      "typeVersion": 1.2
    },
    {
      "id": "cc6b0990-a778-4199-a0fe-80c83dfb9934",
      "name": "Gmail1",
      "type": "n8n-nodes-base.gmail",
      "position": [
        1020,
        -220
      ],
      "webhookId": "[REDACTED_WEBHOOK_ID]",
      "parameters": {
        "labelIds": "={{ $json.output['label ID'] }}",
        "messageId": "={{ $('Gmail').item.json.id }}",
        "operation": "addLabels"
      },
      "credentials": {
        "gmailOAuth2": {
          "id": "[REDACTED_GMAIL_CRED_ID]",
          "name": "Gmail account"
        }
      },
      "typeVersion": 2.1
    },
    {
      "id": "f47e8e55-f634-417f-b62f-60e47af71827",
      "name": "Anthropic Chat Model",
      "type": "@n8n/n8n-nodes-langchain.lmChatAnthropic",
      "position": [
        380,
        160
      ],
      "parameters": {
        "model": {
          "__rl": true,
          "mode": "list",
          "value": "claude-sonnet-4-20250514",
          "cachedResultName": "Claude Sonnet 4"
        },
        "options": {}
      },
      "credentials": {
        "anthropicApi": {
          "id": "[REDACTED_ANTHROPIC_CRED_ID]",
          "name": "Anthropic account 3"
        }
      },
      "typeVersion": 1.3
    },
    {
      "id": "b0ed86c7-5ddd-48dc-aa93-5c4d5c1cacf7",
      "name": "AI Agent",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        500,
        -140
      ],
      "parameters": {
        "text": "Run the task.",
        "options": {
          "systemMessage": "=**Objective:** Analyze the provided email data and classify it with the most appropriate label. **Utilize the capability to check for prior email history (emails sent to or received from the sender by `[REDACTED_EMAIL]`)** to determine if this is a first-time interaction (cold email) or part of an existing relationship. This context is crucial for accurate labeling, especially for distinguishing between Marketing, Notifications, and FYI. Respond with *only* the corresponding Label ID.\n\n**Your Email Address (for context):** `[REDACTED_EMAIL]`\n\n**Key Capabilities (to be used by the system executing this prompt):**\n* **Check Prior Email History:** Before full classification, determine if any prior email correspondence exists between `[REDACTED_EMAIL]` and `{{ $json.from.value[0].address }}`. This check should yield \"Yes\" (history exists) or \"No\" (no history found).\n\n**Input Email Data (from n8n JSON item for the *current* email):**\n* **Sender Email:** `{{ $json.from.value[0].address }}`\n* **Sender Name:** `{{ $json.from.value[0].name }}`\n* **Direct Recipient Emails (To):** `{{ $json.to.value.map(r => r.address).join(', ') }}`\n* **CC Recipient Emails:** `{{ $json.cc.value.map(r => r.address).join(', ') }}` *(Added this based on your JSON, use if populated)*\n* **Subject:** `{{ $json.subject }}`\n* **Body (Plain Text):** `{{ $json.text }}`\n* **Existing Gmail Labels (for current email):** `{{ $json.labelIds.join(', ') }}`\n* **Auto-Submitted Header:** `{{ $json.headers['auto-submitted'] }}`\n* **Original Sender Header (if different from From):** `{{ $json.headers.sender }}`\n* **In-Reply-To Header (for current email):** `{{ $json.headers['in-reply-to'] }}`\n* **References Header (for current email):** `{{ $json.headers.references }}`\n* **List-Unsubscribe Header:** `{{ $json.headers['list-unsubscribe'] }}`\n* **Precedence Header:** `{{ $json.headers.precedence }}` (e.g., \"Bulk\")\n\n**Labels, Descriptions, and Prioritization Logic:**\n\n**Guidance on Using Prior Email History & Unsubscribe Links:**\n* **Prior Email History = \"No\":** Strong indicator of a **cold/unsolicited email**.\n    * If promotional/sales pitch: Likely \"Marketing.\"\n    * If exceptionally personalized & high-value for Trigify: Rare \"To Respond.\"\n* **Prior Email History = \"Yes\":** Indicates an **existing relationship/conversation**.\n    * If it's an update on terms, policies, or service changes from this known entity: Likely \"Notification.\"\n    * If it's a newsletter Max subscribed to: Could be \"FYI\" or \"Marketing\" based on content.\n    * If it's a direct message requiring action: Likely \"To Respond.\"\n* **`List-Unsubscribe` Header or Unsubscribe Links in Body:**\n    * Common in \"Marketing\" emails.\n    * Also present in many legitimate \"Notification\" emails (e.g., service updates, policy changes like the DPA example) and some \"FYI\" newsletters for compliance.\n    * **Therefore, an unsubscribe link alone does not define the category. Consider it alongside Prior Email History and email content/purpose.**\n* **`Precedence: Bulk` Header:** Often indicates mass mailings, common for Marketing, Notifications, and some FYIs.\n\n---\n\n* **To Respond (Label ID: `Label_5151750749488724401`):**\n    * **Primary Criteria:** Requires direct, timely action/reply from `[REDACTED_EMAIL]`.\n    * **Key Indicators (especially if Prior Email History = \"Yes\"):** Direct questions to Max, assigned tasks, requests for info for Trigify, deadlines for Max, part of an active conversation (indicated by `In-Reply-To`/`References`).\n    * **Sales Process Prioritization (Warm Leads/Active Processes - typically Prior Email History = \"Yes\"):** Ongoing, active Trigify sales process discussions.\n    * **High-Value Cold Outreach (Exceptional Cases - Prior Email History = \"No\"):** Highly personalized, strategic opportunity for Trigify requiring Max's personal attention. (Default for cold sales is \"Marketing\").\n\n* **FYI (Label ID: `Label_7518716752151077752`):**\n    * **Primary Criteria:** For Max's awareness; no immediate action/reply required.\n    * **Key Indicators:**\n        * Max is CC'd, primary action for others.\n        * General announcements, *non-promotional* newsletters from *known entities/subscriptions* (Prior Email History = \"Yes\" or sender is clearly a subscribed source) that aren't critical service notifications.\n        * Informational updates within ongoing projects where Max isn't the primary actor.\n    * If an email is a mass mailing (`Precedence: Bulk`, `List-Unsubscribe` present) from a *known entity* (Prior Email History = \"Yes\") and is purely informational without a direct call to action or critical service update, it could be \"FYI.\"\n\n* **Comment (Label ID: `Label_1470120230130814788`):**\n    * Comment/feedback on a document, task, system (e.g., subject \"New comment on...\").\n\n* **Notification (Label ID: `Label_9`):**\n    * **Primary Criteria:** Provides important, often non-promotional, updates or alerts regarding an existing service, account, or system that `[REDACTED_EMAIL]` or Trigify uses or is affected by. Action is typically awareness, potential configuration change, or noting a deadline, rather than a conversational reply.\n    * **Key Indicators:**\n        * **Updates from Known Service Providers (Prior Email History = \"Yes\" is common):**\n            * Changes to Terms of Service, Privacy Policies, Data Processing Agreements (DPAs like the \"Plain\" example).\n            * Critical service availability announcements (outages, maintenance).\n            * Security alerts related to an account Max uses.\n            * Important updates about features or functionality of a service Trigify relies on, which are not primarily marketing new features.\n        * System-generated alerts (e.g., `Auto-Submitted Header` = `auto-generated`) like social media notifications, non-meeting calendar reminders, some financial transaction alerts.\n        * Subject lines may contain: \"Important Update,\" \"Service Notification,\" \"Policy Change,\" \"DPA Update,\" \"Security Alert.\"\n        * Even if a `List-Unsubscribe` link is present (for compliance), if the core content is a critical service/account/legal update from a company Max/Trigify does business with, it's a \"Notification.\"\n        * The `Existing Gmail Labels` might include `CATEGORY_UPDATES`.\n\n* **Meeting Update (Label ID: `Label_7958083858424702466`):**\n    * Update specifically regarding a scheduled meeting (e.g., \"Accepted:\", \"Declined:\", \"Updated Invitation:\", \"Cancelled:\").\n\n* **Marketing (Label ID: `Label_5274550748854913384`):**\n    * **Primary Criteria:** Unsolicited promotional sales pitch, advertisement, or general marketing newsletter, especially if **Prior Email History = \"No.\"**\n    * **Key Indicators:**\n        * **No prior email history found with the `Sender Email`,** AND the email is primarily aimed at selling a product/service or promoting a company/event to Max/Trigify.\n        * Content is generic, focused on features/benefits without strong personalization to Max's known, active projects.\n        * Contains a `List-Unsubscribe Header` AND the content is promotional.\n        * `Existing Gmail Labels` may include `CATEGORY_PROMOTIONS`.\n    * If **Prior Email History = \"Yes\"**: Could still be \"Marketing\" if it's a clearly promotional newsletter/offer from a known contact that doesn't fit \"Notification\" or demand a \"To Respond.\"\n\n---\n\n**Email Data for Classification (Summary):**\n* **Prior Email History with Sender (Yes/No) - From the new capability**\n* Sender Email: `{{ $json.from.value[0].address }}`\n* Sender Name: `{{ $json.from.value[0].name }}`\n* Direct Recipient Emails (To): `{{ $json.to.value.map(r => r.address).join(', ') }}`\n* CC Recipient Emails: `{{ $json.cc.value.map(r => r.address).join(', ') }}`\n* Subject: `{{ $json.subject }}`\n* Body (Plain Text): `{{ $json.text }}`\n* Existing Gmail Labels: `{{ $json.labelIds.join(', ') }}`\n* Auto-Submitted Header: `{{ $json.headers['auto-submitted'] }}`\n* Original Sender Header: `{{ $json.headers.sender }}`\n* In-Reply-To Header: `{{ $json.headers['in-reply-to'] }}`\n* References Header: `{{ $json.headers.references }}`\n* List-Unsubscribe Header: `{{ $json.headers['list-unsubscribe'] }}`\n* Precedence Header: `{{ $json.headers.precedence }}`\n\n**Output:** [Provide ONLY the Label ID here]"
        },
        "promptType": "define",
        "hasOutputParser": true
      },
      "typeVersion": 1.9
    },
    {
      "id": "f4bbdda0-dcd8-41a7-a870-69dfcb747a4f",
      "name": "Get Email",
      "type": "n8n-nodes-base.gmailTool",
      "position": [
        760,
        140
      ],
      "webhookId": "[REDACTED_WEBHOOK_ID]",
      "parameters": {
        "filters": {
          "q": "=from:{{ $fromAI('email') }}"
        },
        "operation": "getAll",
        "returnAll": "={{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Return_All', ``, 'boolean') }}"
      },
      "credentials": {
        "gmailOAuth2": {
          "id": "[REDACTED_GMAIL_CRED_ID_2]",
          "name": "Gmail account 2"
        }
      },
      "typeVersion": 2.1
    },
    {
      "id": "64c3b465-d2be-43cc-ad41-cc1298038640",
      "name": "Check Sent",
      "type": "n8n-nodes-base.gmailTool",
      "position": [
        560,
        180
      ],
      "webhookId": "[REDACTED_WEBHOOK_ID]",
      "parameters": {
        "filters": {
          "q": "=to:{{ $fromAI('email') }}",
          "labelIds": [
            "SENT"
          ]
        },
        "operation": "getAll"
      },
      "credentials": {
        "gmailOAuth2": {
          "id": "[REDACTED_GMAIL_CRED_ID]",
          "name": "Gmail account"
        }
      },
      "typeVersion": 2.1
    }
  ],
  "active": true,
  "pinData": {
    "Gmail Trigger": [
      {
        "json": {
          "To": "[REDACTED_EMAIL]",
          "id": "[REDACTED_MESSAGE_ID]",
          "From": "Coresignal <info@coresignal.com>",
          "labels": [
            {
              "id": "INBOX",
              "name": "INBOX"
            },
            {
              "id": "IMPORTANT",
              "name": "IMPORTANT"
            },
            {
              "id": "CATEGORY_PROMOTIONS",
              "name": "CATEGORY_PROMOTIONS"
            }
          ],
          "Subject": "New MCP server: Direct data access for smarter AI",
          "payload": {
            "mimeType": "multipart/alternative"
          },
          "snippet": "What&#39;s new Empower your AI agent with Coresignal&#39;s data We&#39;re excited to unveil our latest news, the Coresignal MCP server that connects your LLM tools directly to our company, employee,",
          "threadId": "1971723eacc72f08",
          "historyId": "5524551",
          "internalDate": "1748439919000",
          "sizeEstimate": 62610
        }
      },
      {
        "json": {
          "Cc": "",
          "To": "[REDACTED_EMAIL]",
          "id": "[REDACTED_MESSAGE_ID]",
          "From": "Plain <hello@plain.com>",
          "labels": [
            {
              "id": "INBOX",
              "name": "INBOX"
            },
            {
              "id": "IMPORTANT",
              "name": "IMPORTANT"
            },
            {
              "id": "CATEGORY_UPDATES",
              "name": "CATEGORY_UPDATES"
            },
            {
              "id": "UNREAD",
              "name": "UNREAD"
            }
          ],
          "Subject": " Important update to Plain’s Data Processing Agreement (DPA)",
          "payload": {
            "mimeType": "multipart/alternative"
          },
          "snippet": "Hi Max, We&#39;re writing to let you know about an upcoming change to our Data Processing Agreement (DPA). What&#39;s changing? We are simplifying our notification process for subprocessor updates.",
          "threadId": "1971723cfa14c3df",
          "historyId": "5524408",
          "internalDate": "1748439911000",
          "sizeEstimate": 45077
        }
      }
    ]
  },
  "settings": {
    "executionOrder": "v1"
  },
  "versionId": "644d8154-7912-436a-8c05-e8e2f1092993",
  "connections": {
    "Gmail": {
      "main": [
        [
          {
            "node": "AI Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Gmail1": {
      "main": [
        []
      ]
    },
    "AI Agent": {
      "main": [
        [
          {
            "node": "Gmail1",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Get Email": {
      "ai_tool": [
        [
          {
            "node": "AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Check Sent": {
      "ai_tool": [
        [
          {
            "node": "AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Gmail Trigger": {
      "main": [
        [
          {
            "node": "Gmail",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Anthropic Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "AI Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Structured Output Parser": {
      "ai_outputParser": [
        [
          {
            "node": "AI Agent",
            "type": "ai_outputParser",
            "index": 0
          }
        ]
      ]
    }
  }
}