🧑‍🎓 基于AI的语言教师（Telegram、Google表格和GPT-4o）

• •Telegram Bot Token
• •Google Sheets API 凭证
• •OpenAI API Key

{
  "meta": {
    "instanceId": "6a5e68bcca67c4cdb3e0b698d01739aea084e1ec06e551db64aeff43d174cb23",
    "templateCredsSetupCompleted": true
  },
  "nodes": [
    {
      "id": "bc49829b-45f2-4910-9c37-907271982f14",
      "name": "Sticky Note3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -1500,
        -520
      ],
      "parameters": {
        "width": 780,
        "height": 540,
        "content": "### 3. Do you need more details?\nFind a step-by-step guide in this tutorial\n![Guide](https://www.samirsaci.com/content/images/2025/04/Flash-Cards.png)\n[🎥 Watch My Tutorial](https://youtu.be/MQV8wDSug7M)"
      },
      "typeVersion": 1
    },
    {
      "id": "80af5237-9046-4b40-ac7c-167d8e0a490f",
      "name": "AI Agent",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "notes": "Pinyin + Example",
      "position": [
        -2140,
        -140
      ],
      "parameters": {
        "text": "={{ $('Telegram Trigger').item.json.message.text }}",
        "options": {
          "systemMessage": "=# Context\nYou are an AI-powered language tutor designed to help {{ $('Telegram Trigger').item.json.message.chat.first_name }} practice Chinese vocabulary efficiently. \n\n# Role\nYour primary role is to generate interactive Multiple-Choice Questions (MCQs) and evaluate the user's responses.\n\n# Types of Exercises\n- MCQ: Provide an English word and four Chinese answer choices, one correct and three incorrect.\n\n# Rules for MCQ Generation\n1. Select a random **Chinese word** from this list {{ $json.targetLanguage }}\n2. Randomly select **three incorrect Chinese options** from the list or outside the list.\n3. **Do NOT mark the correct answer with ✅** in the question.\n4. Present the question in the following format:\nExample Question Format:\nWhat is the correct translation for \"Warehouse\"?\nA) 运输\nB) 仓库 \nC) 合同\nD) 投标\n5. Ask the user to respond with **A, B, C, or D**.\n\n# Evaluating User Responses:\n1. **Wait for the user's answer. Do NOT assume correctness before checking.**\n2. If the user selects the correct answer:\n- Respond positively: \"Great job! ✅ [Correct Answer] [Correct Answer's Pinyin] means [English Meaning].\"\n3. If the user selects the wrong answer:\n- Provide corrective feedback: \"Oops! ❌ The correct answer was [Correct Answer] ([English Meaning]).\"\n4. If the user provides an **invalid response** (e.g., \"Hello\"), ask them to respond with **A, B, C, or D**.\n\n# Post-Evaluation:\n- After giving feedback, always generate another question. Do not ask the user if he wants another question\n\n# Behavior & Tone\n- Be engaging and encouraging.\n- Ensure clarity in feedback.\n- Guide the user patiently if they provide invalid inputs."
        },
        "promptType": "define",
        "hasOutputParser": true
      },
      "notesInFlow": true,
      "typeVersion": 1.7
    },
    {
      "id": "8b35027e-ec5b-4c3e-9a5b-2780b6c40223",
      "name": "OpenAI Chat Model",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        -2180,
        100
      ],
      "parameters": {
        "model": {
          "__rl": true,
          "mode": "list",
          "value": "gpt-4o-mini"
        },
        "options": {}
      },
      "typeVersion": 1.2
    },
    {
      "id": "688d6882-4930-407d-bf58-5f6add8eb159",
      "name": "Simple Memory",
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "position": [
        -2000,
        140
      ],
      "parameters": {
        "sessionKey": "={{ $('Telegram Trigger').item.json.message.chat.id }}",
        "sessionIdType": "customKey"
      },
      "typeVersion": 1.3
    },
    {
      "id": "33f4a062-73f9-4a99-abca-1184ef2c2a41",
      "name": "Telegram Trigger",
      "type": "n8n-nodes-base.telegramTrigger",
      "position": [
        -2960,
        -140
      ],
      "webhookId": "88179da7-9927-4bdc-8bd7-78022810b48e",
      "parameters": {
        "updates": [
          "message"
        ],
        "additionalFields": {}
      },
      "notesInFlow": true,
      "typeVersion": 1.1
    },
    {
      "id": "af385807-d024-477e-9a42-c195043e95da",
      "name": "Retrive Vocabulary",
      "type": "n8n-nodes-base.googleSheets",
      "position": [
        -2700,
        -140
      ],
      "parameters": {
        "options": {},
        "sheetName": {
          "__rl": true,
          "mode": "list",
          "value": 0,
          "cachedResultUrl": "=",
          "cachedResultName": "="
        },
        "documentId": {
          "__rl": true,
          "mode": "list",
          "value": "=",
          "cachedResultUrl": "=",
          "cachedResultName": "="
        }
      },
      "notesInFlow": true,
      "typeVersion": 4.5
    },
    {
      "id": "3ab67ca5-9839-4fa6-bfc1-4dbbaf5593fc",
      "name": "Sticky Note1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -3000,
        -520
      ],
      "parameters": {
        "color": 7,
        "width": 680,
        "height": 540,
        "content": "### 1. Workflow Trigger with Telegram Message\n1. The workflow is triggered by a user message. \n2. The second node retrieves the vocabulary list from a Google Sheet.\n3. The third node combines all the words in Chinese and English in two distinctive lists.\n\n#### How to setup?\n- **Telegram Node:** set up your telegram bot credentials\n[Learn more about the Telegram Trigger Node](https://docs.n8n.io/integrations/builtin/trigger-nodes/n8n-nodes-base.telegramtrigger/)\n- **Retrieve Vocabulary from a Google Sheet Node**:\n   1. Add your Google Sheet API credentials to access the Google Sheet file\n   2. Select the file using the list, an URL or an ID\n   3. Select the sheet in which you have stored your vocabulary list\n  [Learn more about the Google Sheet Node](https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.googlesheets)\n"
      },
      "typeVersion": 1
    },
    {
      "id": "740a2d04-46fe-41f1-b887-f88f3e23c50d",
      "name": "Sticky Note",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -2300,
        -520
      ],
      "parameters": {
        "color": 7,
        "width": 760,
        "height": 780,
        "content": "### 2. Conversational AI Agent\nThe AI agent will take as inputs the two vocabulary lists and user's message to asks questions and process answers. Conversations are recorded by chat id; each user has its own conversation with the bot.\n\n#### How to setup?\n- **Telegram Nodes:** set up your telegram bot credentials\n[Learn more about the Telegram Trigger Node](https://docs.n8n.io/integrations/builtin/trigger-nodes/n8n-nodes-base.telegramtrigger/)\n- **AI Agent with the Chat Model**:\n   1. Add a chat model with the required credentials *(Example: Open AI 4o-mini)*\n   2. Adapt the system prompt with the **target learning language** and the format of the question you want to have.\n"
      },
      "typeVersion": 1
    },
    {
      "id": "e92a55dc-6d9d-4008-bb40-72a7f2dd470c",
      "name": "Aggregate Vocabulary Lists",
      "type": "n8n-nodes-base.aggregate",
      "position": [
        -2460,
        -140
      ],
      "parameters": {
        "options": {},
        "fieldsToAggregate": {
          "fieldToAggregate": [
            {
              "renameField": true,
              "outputFieldName": "initialLanguage",
              "fieldToAggregate": "initialText"
            },
            {
              "renameField": true,
              "outputFieldName": "targetLanguage",
              "fieldToAggregate": "translatedText"
            }
          ]
        }
      },
      "typeVersion": 1
    },
    {
      "id": "18b29677-cfc0-4817-9321-35090a3fda2e",
      "name": "Answer to the User",
      "type": "n8n-nodes-base.telegram",
      "position": [
        -1740,
        -140
      ],
      "webhookId": "=",
      "parameters": {
        "text": "={{ $json.output }}",
        "chatId": "={{ $('Telegram Trigger').item.json.message.chat.id }}",
        "additionalFields": {
          "appendAttribution": false
        }
      },
      "notesInFlow": true,
      "typeVersion": 1.2
    }
  ],
  "pinData": {},
  "connections": {
    "80af5237-9046-4b40-ac7c-167d8e0a490f": {
      "main": [
        [
          {
            "node": "18b29677-cfc0-4817-9321-35090a3fda2e",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "688d6882-4930-407d-bf58-5f6add8eb159": {
      "ai_memory": [
        [
          {
            "node": "80af5237-9046-4b40-ac7c-167d8e0a490f",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "33f4a062-73f9-4a99-abca-1184ef2c2a41": {
      "main": [
        [
          {
            "node": "af385807-d024-477e-9a42-c195043e95da",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "8b35027e-ec5b-4c3e-9a5b-2780b6c40223": {
      "ai_languageModel": [
        [
          {
            "node": "80af5237-9046-4b40-ac7c-167d8e0a490f",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "af385807-d024-477e-9a42-c195043e95da": {
      "main": [
        [
          {
            "node": "e92a55dc-6d9d-4008-bb40-72a7f2dd470c",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "e92a55dc-6d9d-4008-bb40-72a7f2dd470c": {
      "main": [
        [
          {
            "node": "80af5237-9046-4b40-ac7c-167d8e0a490f",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}