用 n8n 自动分析、分类并总结 Gmail 邮件（OpenAI RAG + Google Sheets）

Gmail 新邮件触发 → 提取清洗 → RAG 向量检索分类 → Google Sheets 存储 → 自学习分类法更新 → Telegram 音频摘要通知

给谁

需要自动化处理大量 Gmail 邮件、并按自定义分类法归档的销售/客服/运营人员

你会得到

新邮件到达后自动提取内容，通过 RAG 检索历史分类示例进行 AI 分类与摘要，结果写入 Google Sheets，同时生成音频摘要推送至 Telegram；遇到新类别时自动扩展分类法并通知管理员

你需要

n8n · gmail · google-sheets · telegram

步骤

1. 配置触发器与凭证

   在 n8n 中创建以下凭证：Gmail OAuth2（用于触发和读取邮件）、Google Sheets OAuth2（用于存储和分类法数据集）、OpenAI API（用于聊天、嵌入和音频生成）、Telegram API（用于通知）。添加 Gmail Trigger 节点，设置为每分钟轮询新邮件。添加 Fetch full Gmail message 节点，使用 messageId 获取完整邮件内容。

2. 提取并规范化邮件字段

   添加 Normalize email fields（Edit Fields/Set）节点，从原始邮件中提取并映射以下字段：id、date、sender（from.value[0].name）、email（from.value[0].address）、subject、content（使用 html.removeTags() 去除 HTML 标签）。确保输出为纯文本，供后续 AI 节点处理。

3. 准备 RAG 分类法知识库（首次手动运行）

   在 Google Sheets 中创建两个工作表：Sheet 1「Email Log」含列 date、email、name、subject、summarized text、category、subcategory、keywords；Sheet 2「Tagging Samples」含列 id、subject、email_text、category、subcategory、keywords。添加 Get row(s) in sheet 节点读取 Tagging Samples 全表；添加 Prepare taxonomy documents 节点将 subject 与 email_text 拼接为文档；添加 Create embeddings (OpenAI) 节点生成向量嵌入；添加 Vector store — taxonomy index 节点以 insert 模式写入内存向量库（memoryKey=email_tagging）。添加 Vector store retriever (RAG tool) 节点以 retrieve-as-tool 模式供 Agent 调用（topK=1）。首次需手动点击「Execute workflow」运行此分支以填充向量存储。

4. 配置 AI 分析 Agent（RAG 分类与摘要）

   添加 Analyze email with RAG agent 节点（AI Agent），连接 OpenAI Chat Model（gpt-4.1-mini）作为语言模型，连接 Vector store retriever 作为工具。在 Agent 的 text 字段输入邮件变量，在 options.systemMessage 粘贴系统提示词。该提示词要求模型：用向量存储检索到的示例作为分类法唯一真相来源，优先使用已有类别，仅在无合理匹配时创建新类别；输出严格 JSON，附加 summary、short_message、category、subcategory、keywords、confidence、new_category、new_subcategory 字段；并在最终输出前做内部校验。

   运行提示词 · 来源原文。AI Agent 节点的 text 输入字段（User Message），需保留 n8n 表达式语法

   复制提示词

   =sender: {{ $json.sender }}
   email: {{ $json.email }}
   date: {{ $json.date }}
   subject: {{ $json.subject }}
   content: {{ $json.content }}

   运行提示词 · 来源原文。AI Agent 节点的 options.systemMessage 字段（System Prompt）

   复制提示词

   =You are an email analysis, classification, and summarization assistant operating inside an automated workflow.
   
   You will receive ONE incoming email object as input.
   
   Your task is to analyze the email and RETURN THE SAME OBJECT with additional structured fields appended.
   
   --------------------------------------------------
   PRIMARY OBJECTIVES
   --------------------------------------------------
   
   1. Summarize the email content in ONE short paragraph (maximum 3 sentences).
   
   2. Classify the email using semantic context retrieved from "Simple Vector Store1".
   
   3. The retrieved examples represent the OFFICIAL taxonomy and are the PRIMARY SOURCE OF TRUTH for categories and subcategories.
   
   4. Extract up to 5 meaningful keywords representing the main topics of the email.
   
   5. Provide a confidence score between 0 and 1 indicating classification certainty.
   
   --------------------------------------------------
   CLASSIFICATION RULES (STRICT)
   --------------------------------------------------
   
   - ALWAYS prefer categories and subcategories found in retrieved examples.
   - DO NOT invent new categories or subcategories unless no reasonable semantic match exists.
   - Retrieved examples have higher authority than general model knowledge.
   - If similarity is weak, still choose the closest existing category and LOWER the confidence score.
   - Only create a new category or subcategory when NO retrieved example reasonably matches the email topic.
   
   Decision priority order:
   1) Retrieved vector store examples (highest priority)
   2) Semantic meaning of the email
   3) General reasoning (last resort)
   
   --------------------------------------------------
   TAXONOMY CONSISTENCY CHECK (MANDATORY)
   --------------------------------------------------
   
   After selecting a category and subcategory:
   
   1. Compare the selected category against ALL categories present in the retrieved examples from "Simple Vector Store1".
   
   2. If the selected category DOES NOT appear in retrieved examples from knowledge_base and you generated new one:
      set new_category = true
      otherwise set new_category = false.
   
   3. If the category exists but the selected subcategory DOES NOT appear in retrieved examples from knowledge_base and you generated new one:
      set new_subcategory = true
      otherwise set new_subcategory = false.
   
   4. This decision MUST be based ONLY on retrieved examples, not on general knowledge.
   
   --------------------------------------------------
   CONFIDENCE SCORING
   --------------------------------------------------
   
   0.90–1.00 → strong semantic match with retrieved examples
   0.70–0.89 → clear topic with minor ambiguity
   0.50–0.69 → weak similarity but reasonable classification
   0.00–0.49 → uncertain classification
   
   Lower confidence when retrieved similarity is weak or ambiguous.
   
   --------------------------------------------------
   OUTPUT REQUIREMENTS (CRITICAL)
   --------------------------------------------------
   
   - Output MUST be valid JSON ONLY.
   - DO NOT include markdown, explanations, comments, or extra text.
   - DO NOT remove, rename, or modify any existing input fields.
   - Preserve all original fields exactly as received.
   - Append ONLY the following fields:
   
   {
     "summary": "short paragraph summary",
     "short_message": "1 long line summary",
     "category": "main category",
     "subcategory": "specific subcategory",
     "keywords": ["keyword1","keyword2","keyword3"],
     "confidence": 0.0,
     "new_category": false,
     "new_subcategory": false
   }
   
   --------------------------------------------------
   NEW CATEGORY FLAGS
   --------------------------------------------------
   
   - new_category = true ONLY if the selected category is NOT present in retrieved examples.
   - new_subcategory = true ONLY if the category exists but the subcategory is NOT present in retrieved examples.
   
   --------------------------------------------------
   FINAL VALIDATION STEP (MANDATORY)
   --------------------------------------------------
   
   Before responding, internally verify:
   
   ✓ Output is valid JSON
   ✓ No missing commas or trailing text
   ✓ All original input fields remain unchanged
   ✓ Exactly the required fields were appended
   ✓ No additional fields were added
   ✓ Boolean fields contain true or false values only
   ✓ confidence is a number between 0 and 1
   
   If validation fails, correct the output before responding.

5. 解析结果并存储到 Google Sheets

   添加 Parse AI analysis result（Edit Fields/Set）节点，使用 parseJson() 解析 Agent 的 JSON 输出，映射为 sender、email、date、subject、content（summary）、short message、category、subcategory、keyword（keywords.join()）、confidence、new category、new subcategory。随后添加 Store processed email（Google Sheets）节点，以 append 模式写入 Email Log 工作表，字段映射与步骤 3 中 Sheet 1 的列结构一致。

6. 自学习分类法更新与管理员通知

   添加 Check for new taxonomy items（If）节点，判断 new category 或 new subcategory 是否为 "true"。若为真，走「true」分支：添加 Load taxonomy dataset 读取现有 Tagging Samples；添加 Get next taxonomy ID（Code 节点，Python）计算最大 id；添加 Append new taxonomy entry 将新类别/子类别写入 Tagging Samples；添加 Notify admin about taxonomy update（Telegram）发送文字通知。

   运行提示词 · 来源原文。Telegram 节点 Notify admin about taxonomy update 的 text 字段

   复制提示词

   =New category/subcategory added:
   {{ $json.category }} / {{ $json.subcategory }}

7. 生成音频摘要并推送 Telegram

   在 Check for new taxonomy items 的「false」分支（或直接并行）添加 Generate audio summary（OpenAI）节点，resource 设为 audio，input 取 Store processed email 的 summarized text。再添加 Send audio summary to Telegram 节点，operation 设为 sendAudio，binaryData=true，caption 包含发件人、格式化日期、主题，fileName 设为 "email bite"。chatId 需替换为实际 Telegram 用户或群组 ID。

你正在搭建一个 n8n 工作流，用于自动分析、分类和总结 Gmail 邮件。请按以下顺序确认并执行：
1. 检查你是否已具备：n8n 实例、Google 账号（Gmail + Sheets）、OpenAI API key、Telegram Bot token 与 chat ID。
2. 在 n8n 中创建四个凭证：Gmail OAuth2、Google Sheets OAuth2、OpenAI API、Telegram API。
3. 按步骤说明在 Google Sheets 中创建 Email Log 和 Tagging Samples 两个工作表，并填入初始分类示例。
4. 导入或按 workflow.json 重建节点连线，特别注意：AI Agent 的 systemMessage 和 text 字段必须逐字复制；Vector store 的 memoryKey 统一为 email_tagging；首次运行需手动执行以填充向量存储。
5. 将 Telegram 节点中的 [YOUR_CHAT_ID] 替换为真实 ID。
6. 激活工作流后，发送一封测试邮件，验证：邮件是否出现在 Email Log、类别是否正确、Telegram 是否收到音频摘要。