Voice RAG Chatbot with ElevenLabs and OpenAI

工作流概述

这是一个包含23个节点的复杂工作流,主要用于自动化处理各种任务。

工作流源代码

下载
{
  "id": "ibiHg6umCqvcTF4g",
  "meta": {
    "instanceId": "a4bfc93e975ca233ac45ed7c9227d84cf5a2329310525917adaf3312e10d5462",
    "templateCredsSetupCompleted": true
  },
  "name": "Voice RAG Chatbot with ElevenLabs and OpenAI",
  "tags": [],
  "nodes": [
    {
      "id": "5898da57-38b0-4d29-af25-fe029cda7c4a",
      "name": "AI Agent",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        -180,
        800
      ],
      "parameters": {
        "text": "={{ $json.body.question }}",
        "options": {},
        "promptType": "define"
      },
      "typeVersion": 1.7
    },
    {
      "id": "81bbedb6-5a07-4977-a68f-2bdc75b17aba",
      "name": "Vector Store Tool",
      "type": "@n8n/n8n-nodes-langchain.toolVectorStore",
      "position": [
        20,
        1040
      ],
      "parameters": {
        "name": "company",
        "description": "Risponde alle domande relative a ciò che ti viene chiesto"
      },
      "typeVersion": 1
    },
    {
      "id": "fd021f6c-248d-41f4-a4f9-651e70692327",
      "name": "Qdrant Vector Store",
      "type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
      "position": [
        -140,
        1300
      ],
      "parameters": {
        "options": {},
        "qdrantCollection": {
          "__rl": true,
          "mode": "id",
          "value": "=COLLECTION"
        }
      },
      "credentials": {
        "qdrantApi": {
          "id": "iyQ6MQiVaF3VMBmt",
          "name": "QdrantApi account"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "84aca7bb-4812-498f-b319-88831e4ca412",
      "name": "Embeddings OpenAI",
      "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
      "position": [
        -140,
        1460
      ],
      "parameters": {
        "options": {}
      },
      "credentials": {
        "openAiApi": {
          "id": "CDX6QM4gLYanh0P4",
          "name": "OpenAi account"
        }
      },
      "typeVersion": 1.1
    },
    {
      "id": "82e430db-2ad7-427d-bcf9-6aa226253d18",
      "name": "Sticky Note4",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -760,
        520
      ],
      "parameters": {
        "color": 5,
        "width": 1400,
        "height": 240,
        "content": "# STEP 4

## RAG System

Click on \"test workflow\" on n8n and \"Test AI agent\" on ElevenLabs. If everything is configured correctly, when you ask a question to the agent, the webhook on n8n is activated with the \"question\" field in the body filled with the question asked to the voice agent.

The AI ​​Agent will extract the information from the vector database, send it to the model to create the response which will be sent via the response webhook to ElevenLabs which will transform it into voice"
      },
      "typeVersion": 1
    },
    {
      "id": "6a19e9fa-50fa-4d51-ba41-d03c999e4649",
      "name": "Sticky Note",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -780,
        -880
      ],
      "parameters": {
        "color": 3,
        "width": 1420,
        "height": 360,
        "content": "# STEP 1

## Create an Agent on ElevenLabs 
- Create an agent on ElevenLabs (eg. test_n8n)
- Add \"First message\" (eg. Hi, Can I help you?)
- Add the \"System Prompt\" message... eg:
'You are the waiter of \"Pizzeria da Michele\" in Verona. If you are asked a question, use the tool \"test_chatbot_elevenlabs\". When you receive the answer from \"test_chatbot_elevenlabs\" answer the user clearly and precisely.'
- In Tools add a Webhook called eg. \"test_chatbot_elevenlabs\" and add the following description:
'You are the waiter. Answer the questions asked and store them in the question field.'
- Add the n8n webhook URL (method POST)
- Enable \"Body Parameters\" and insert in the description \"Ask the user the question to ask the place.\", then in the \"Properties\" add a data type string called \"question\", value type \"LLM Prompt\" and description \"user question\""
      },
      "typeVersion": 1
    },
    {
      "id": "ec053ee7-3a4a-4697-a08c-5645810d23f0",
      "name": "When clicking ‘Test workflow’",
      "type": "n8n-nodes-base.manualTrigger",
      "position": [
        -740,
        -200
      ],
      "parameters": {},
      "typeVersion": 1
    },
    {
      "id": "3e71e40c-a5cc-40cf-a159-aeedc97c47d1",
      "name": "Create collection",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        -440,
        -340
      ],
      "parameters": {
        "url": "https://QDRANTURL/collections/COLLECTION",
        "method": "POST",
        "options": {},
        "jsonBody": "{
  \"filter\": {}
}",
        "sendBody": true,
        "sendHeaders": true,
        "specifyBody": "json",
        "authentication": "genericCredentialType",
        "genericAuthType": "httpHeaderAuth",
        "headerParameters": {
          "parameters": [
            {
              "name": "Content-Type",
              "value": "application/json"
            }
          ]
        }
      },
      "credentials": {
        "httpHeaderAuth": {
          "id": "qhny6r5ql9wwotpn",
          "name": "Qdrant API (Hetzner)"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "240283fc-50ec-475c-bd24-e6d0a367c10c",
      "name": "Refresh collection",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        -440,
        -80
      ],
      "parameters": {
        "url": "https://QDRANTURL/collections/COLLECTION/points/delete",
        "method": "POST",
        "options": {},
        "jsonBody": "{
  \"filter\": {}
}",
        "sendBody": true,
        "sendHeaders": true,
        "specifyBody": "json",
        "authentication": "genericCredentialType",
        "genericAuthType": "httpHeaderAuth",
        "headerParameters": {
          "parameters": [
            {
              "name": "Content-Type",
              "value": "application/json"
            }
          ]
        }
      },
      "credentials": {
        "httpHeaderAuth": {
          "id": "qhny6r5ql9wwotpn",
          "name": "Qdrant API (Hetzner)"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "7d10fda0-c6ab-4bf5-b73e-b93a84937eff",
      "name": "Get folder",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        -220,
        -80
      ],
      "parameters": {
        "filter": {
          "driveId": {
            "__rl": true,
            "mode": "list",
            "value": "My Drive",
            "cachedResultUrl": "https://drive.google.com/drive/my-drive",
            "cachedResultName": "My Drive"
          },
          "folderId": {
            "__rl": true,
            "mode": "id",
            "value": "=test-whatsapp"
          }
        },
        "options": {},
        "resource": "fileFolder"
      },
      "credentials": {
        "googleDriveOAuth2Api": {
          "id": "HEy5EuZkgPZVEa9w",
          "name": "Google Drive account"
        }
      },
      "typeVersion": 3
    },
    {
      "id": "c5761ad2-e66f-4d65-b653-0e89ea017f17",
      "name": "Download Files",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        0,
        -80
      ],
      "parameters": {
        "fileId": {
          "__rl": true,
          "mode": "id",
          "value": "={{ $json.id }}"
        },
        "options": {
          "googleFileConversion": {
            "conversion": {
              "docsToFormat": "text/plain"
            }
          }
        },
        "operation": "download"
      },
      "credentials": {
        "googleDriveOAuth2Api": {
          "id": "HEy5EuZkgPZVEa9w",
          "name": "Google Drive account"
        }
      },
      "typeVersion": 3
    },
    {
      "id": "1f031a11-8ef3-4392-a7db-9bca00840b8f",
      "name": "Default Data Loader",
      "type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
      "position": [
        380,
        120
      ],
      "parameters": {
        "options": {},
        "dataType": "binary"
      },
      "typeVersion": 1
    },
    {
      "id": "7f614392-7bc7-408c-8108-f289a81d5cf6",
      "name": "Token Splitter",
      "type": "@n8n/n8n-nodes-langchain.textSplitterTokenSplitter",
      "position": [
        360,
        280
      ],
      "parameters": {
        "chunkSize": 300,
        "chunkOverlap": 30
      },
      "typeVersion": 1
    },
    {
      "id": "648c5b3d-37a8-4a89-b88c-38e1863f09dc",
      "name": "Sticky Note3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -240,
        -400
      ],
      "parameters": {
        "color": 6,
        "width": 880,
        "height": 220,
        "content": "# STEP 2

## Create Qdrant Collection
Change:
- QDRANTURL
- COLLECTION"
      },
      "typeVersion": 1
    },
    {
      "id": "a6c50f3c-3c73-464e-9bdc-49de96401c1b",
      "name": "Qdrant Vector Store1",
      "type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
      "position": [
        240,
        -80
      ],
      "parameters": {
        "mode": "insert",
        "options": {},
        "qdrantCollection": {
          "__rl": true,
          "mode": "id",
          "value": "=COLLECTION"
        }
      },
      "credentials": {
        "qdrantApi": {
          "id": "iyQ6MQiVaF3VMBmt",
          "name": "QdrantApi account"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "7e19ac49-4d90-4258-bd44-7ca4ffa0128a",
      "name": "Embeddings OpenAI1",
      "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
      "position": [
        220,
        120
      ],
      "parameters": {
        "options": {}
      },
      "credentials": {
        "openAiApi": {
          "id": "CDX6QM4gLYanh0P4",
          "name": "OpenAi account"
        }
      },
      "typeVersion": 1.1
    },
    {
      "id": "bfa104a2-1f9c-4200-ae7b-4659894c1e6f",
      "name": "Sticky Note5",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -460,
        -140
      ],
      "parameters": {
        "color": 4,
        "width": 620,
        "height": 400,
        "content": "# STEP 3












## Documents vectorization with Qdrant and Google Drive
Change:
- QDRANTURL
- COLLECTION"
      },
      "typeVersion": 1
    },
    {
      "id": "a148ffcf-335f-455d-8509-d98c711ed740",
      "name": "Respond to ElevenLabs",
      "type": "n8n-nodes-base.respondToWebhook",
      "position": [
        380,
        800
      ],
      "parameters": {
        "options": {}
      },
      "typeVersion": 1.1
    },
    {
      "id": "5d19f73a-b8e8-4e75-8f67-836180597572",
      "name": "OpenAI",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        -300,
        1040
      ],
      "parameters": {
        "options": {}
      },
      "credentials": {
        "openAiApi": {
          "id": "CDX6QM4gLYanh0P4",
          "name": "OpenAi account"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "802b76e1-3f3e-490c-9e3b-65dc5b28d906",
      "name": "Listen",
      "type": "n8n-nodes-base.webhook",
      "position": [
        -700,
        800
      ],
      "webhookId": "e9f611eb-a8dd-4520-8d24-9f36deaca528",
      "parameters": {
        "path": "test_voice_message_elevenlabs",
        "options": {},
        "httpMethod": "POST",
        "responseMode": "responseNode"
      },
      "typeVersion": 2
    },
    {
      "id": "bdc55a38-1d4b-48fe-bbd8-29bf1afd954a",
      "name": "Window Buffer Memory",
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "position": [
        -140,
        1040
      ],
      "parameters": {},
      "typeVersion": 1.3
    },
    {
      "id": "2d5dd8cb-81eb-41bc-af53-b894e69e530c",
      "name": "OpenAI Chat Model",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        200,
        1320
      ],
      "parameters": {
        "options": {}
      },
      "credentials": {
        "openAiApi": {
          "id": "CDX6QM4gLYanh0P4",
          "name": "OpenAi account"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "92d04432-1dbb-4d79-9edc-42378aee1c53",
      "name": "Sticky Note6",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -760,
        1620
      ],
      "parameters": {
        "color": 7,
        "width": 1400,
        "height": 240,
        "content": "# STEP 5

## Add Widget

Add the widget to your business website by replacing AGENT_ID with the agent id you created on ElevenLabs

<elevenlabs-convai agent-id=\"AGENT_ID\"></elevenlabs-convai><script src=\"https://elevenlabs.io/convai-widget/index.js\" async type=\"text/javascript\"></script>"
      },
      "typeVersion": 1
    }
  ],
  "active": false,
  "pinData": {},
  "settings": {
    "executionOrder": "v1"
  },
  "versionId": "6738abfe-e626-488d-a00b-81021cb04aaf",
  "connections": {
    "Listen": {
      "main": [
        [
          {
            "node": "AI Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "OpenAI": {
      "ai_languageModel": [
        [
          {
            "node": "AI Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "AI Agent": {
      "main": [
        [
          {
            "node": "Respond to ElevenLabs",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Get folder": {
      "main": [
        [
          {
            "node": "Download Files",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Download Files": {
      "main": [
        [
          {
            "node": "Qdrant Vector Store1",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Token Splitter": {
      "ai_textSplitter": [
        [
          {
            "node": "Default Data Loader",
            "type": "ai_textSplitter",
            "index": 0
          }
        ]
      ]
    },
    "Embeddings OpenAI": {
      "ai_embedding": [
        [
          {
            "node": "Qdrant Vector Store",
            "type": "ai_embedding",
            "index": 0
          }
        ]
      ]
    },
    "OpenAI Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "Vector Store Tool",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Vector Store Tool": {
      "ai_tool": [
        [
          {
            "node": "AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Embeddings OpenAI1": {
      "ai_embedding": [
        [
          {
            "node": "Qdrant Vector Store1",
            "type": "ai_embedding",
            "index": 0
          }
        ]
      ]
    },
    "Refresh collection": {
      "main": [
        [
          {
            "node": "Get folder",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Default Data Loader": {
      "ai_document": [
        [
          {
            "node": "Qdrant Vector Store1",
            "type": "ai_document",
            "index": 0
          }
        ]
      ]
    },
    "Qdrant Vector Store": {
      "ai_vectorStore": [
        [
          {
            "node": "Vector Store Tool",
            "type": "ai_vectorStore",
            "index": 0
          }
        ]
      ]
    },
    "Window Buffer Memory": {
      "ai_memory": [
        [
          {
            "node": "AI Agent",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "When clicking ‘Test workflow’": {
      "main": [
        [
          {
            "node": "Create collection",
            "type": "main",
            "index": 0
          },
          {
            "node": "Refresh collection",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}

功能特点

  • 自动检测新邮件
  • AI智能内容分析
  • 自定义分类规则
  • 批量处理能力
  • 详细的处理日志

技术分析

节点类型及作用

  • @N8N/N8N Nodes Langchain.Agent
  • @N8N/N8N Nodes Langchain.Toolvectorstore
  • @N8N/N8N Nodes Langchain.Vectorstoreqdrant
  • @N8N/N8N Nodes Langchain.Embeddingsopenai
  • Stickynote

复杂度评估

配置难度:
★★★★☆
维护难度:
★★☆☆☆
扩展性:
★★★★☆

实施指南

前置条件

  • 有效的Gmail账户
  • n8n平台访问权限
  • Google API凭证
  • AI分类服务订阅

配置步骤

  1. 在n8n中导入工作流JSON文件
  2. 配置Gmail节点的认证信息
  3. 设置AI分类器的API密钥
  4. 自定义分类规则和标签映射
  5. 测试工作流执行
  6. 配置定时触发器(可选)

关键参数

参数名称 默认值 说明
maxEmails 50 单次处理的最大邮件数量
confidenceThreshold 0.8 分类置信度阈值
autoLabel true 是否自动添加标签

最佳实践

优化建议

  • 定期更新AI分类模型以提高准确性
  • 根据邮件量调整处理批次大小
  • 设置合理的分类置信度阈值
  • 定期清理过期的分类规则

安全注意事项

  • 妥善保管API密钥和认证信息
  • 限制工作流的访问权限
  • 定期审查处理日志
  • 启用双因素认证保护Gmail账户

性能优化

  • 使用增量处理减少重复工作
  • 缓存频繁访问的数据
  • 并行处理多个邮件分类任务
  • 监控系统资源使用情况

故障排除

常见问题

邮件未被正确分类

检查AI分类器的置信度阈值设置,适当降低阈值或更新训练数据。

Gmail认证失败

确认Google API凭证有效且具有正确的权限范围,重新进行OAuth授权。

调试技巧

  • 启用详细日志记录查看每个步骤的执行情况
  • 使用测试邮件验证分类逻辑
  • 检查网络连接和API服务状态
  • 逐步执行工作流定位问题节点

错误处理

工作流包含以下错误处理机制:

  • 网络超时自动重试(最多3次)
  • API错误记录和告警
  • 处理失败邮件的隔离机制
  • 异常情况下的回滚操作