8
n8n 中文网amn8n.com

使用Vision API和LLM实现自动化发票收集与数据提取

高级

这是一个Invoice Processing, Multimodal AI领域的自动化工作流，包含 17 个节点。主要使用 Set, Code, Telegram, FormTrigger, GoogleDrive 等节点。 使用Google Vision OCR、Gemini LLM和Google Sheets提取和结构化发票数据

前置要求
  • •Telegram Bot Token
  • •Google Drive API 凭证
  • •可能需要目标 API 的认证凭证
  • •Google Sheets API 凭证

使用的节点 (17)

Set
Code
Telegram
StickyNote
FormTrigger
GoogleDrive
HttpRequest
GoogleSheets
ChainLlm
LmChatOpenRouter
OutputParserStructured

分类

发票处理
多模态 AI
工作流预览
可视化展示节点连接关系，支持缩放和平移
React Flow
导出工作流
复制以下 JSON 配置到 n8n 导入，即可使用此工作流
{
  "id": "weV4UB3UBRHEfe1k",
  "meta": {
    "templateCredsSetupCompleted": true
  },
  "name": "使用 Vision API 和 LLM 实现自动化发票收集与数据提取",
  "tags": [],
  "nodes": [
    {
      "id": "2b378a23-08b9-4b7e-b1c5-114087802a71",
      "name": "便签1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -340,
        -480
      ],
      "parameters": {
        "width": 580,
        "height": 1200,
        "content": "# 使用 Vision API 和 LLM 实现自动化发票收集与数据提取"
      },
      "typeVersion": 1
    },
    {
      "id": "e4ac27f9-6db0-489f-bd7f-033fd792fc7c",
      "name": "设置 Vision API",
      "type": "n8n-nodes-base.set",
      "position": [
        1120,
        20
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "586cf442-13be-4996-9824-366e20ab864e",
              "name": "visionAPI",
              "type": "string",
              "value": "YOUR_VISION_API_KEY_HERE"
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "2e81ab2f-5f8f-4676-bfbf-9753bad4369e",
      "name": "基础 LLM 链",
      "type": "@n8n/n8n-nodes-langchain.chainLlm",
      "position": [
        1540,
        20
      ],
      "parameters": {
        "text": "={{ $json.responses[0].fullTextAnnotation.text }}",
        "batching": {},
        "messages": {
          "messageValues": [
            {
              "message": "=You are a professional accountant experienced in preparing financial reports and classifying expenses into appropriate categories.\nYou will receive input data extracted via OCR that needs to be parsed and structured.\n\nInstructions:\n\nOutput must be in the same language as the input.\nParse and return the data in this column order:\nDATE : Extract in YYYY-MM-DD format, if not found use current date\nNO VOUCHER : Extract invoice or voucher number, if not found return not found\nTRANSACTION DETAIL : Short summary of products or services, if not found return not found\nVENDOR : Store or vendor name, if not found return not found\nVALUE : Transaction total amount as a number without currency symbol, if not found return 0\n\nMissing text fields : not found\nMissing numeric values : 0\nOutput only the structured data, no extra explanations"
            }
          ]
        },
        "promptType": "define",
        "hasOutputParser": true
      },
      "typeVersion": 1.7
    },
    {
      "id": "d0ca7089-c557-48d8-a854-310efbc5b5be",
      "name": "OpenRouter 聊天模型",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenRouter",
      "position": [
        1540,
        220
      ],
      "parameters": {
        "model": "google/gemini-2.0-flash-exp:free",
        "options": {}
      },
      "typeVersion": 1
    },
    {
      "id": "22bbee6d-a897-4f2a-b0c9-06b7574ebf8e",
      "name": "结构化输出解析器",
      "type": "@n8n/n8n-nodes-langchain.outputParserStructured",
      "position": [
        1700,
        220
      ],
      "parameters": {
        "jsonSchemaExample": "{\n  \"DATE\": \"2025-07-07\",\n  \"NO VOUCHER\": \"INV-001234\",\n  \"TRANSACTION DETAIL\": \"Product summary\",\n  \"VENDOR\": \"Vendor Name or Shop Name\",\n  \"VALUE\": 0\n}"
      },
      "typeVersion": 1.3
    },
    {
      "id": "8a108ae0-3061-49a9-aef1-c426f078cca1",
      "name": "下载文件",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        740,
        20
      ],
      "parameters": {
        "fileId": {
          "__rl": true,
          "mode": "id",
          "value": "={{ $json.id }}"
        },
        "options": {},
        "operation": "download"
      },
      "typeVersion": 3
    },
    {
      "id": "0a96b787-f8e4-408c-80d8-dc2a868c93e4",
      "name": "HTTP 请求",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        1320,
        20
      ],
      "parameters": {
        "url": "=https://vision.googleapis.com/v1/images:annotate?key={{ $json.visionAPI }}",
        "method": "POST",
        "options": {},
        "jsonBody": "={\n  \"requests\": [\n    {\n      \"image\": {\n        \"content\": \"{{ $('Code').item.json.base64 }}\"\n      },\n      \"features\": [\n        {\n          \"type\": \"TEXT_DETECTION\"\n        }\n      ]\n    }\n  ]\n}\n",
        "sendBody": true,
        "specifyBody": "json"
      },
      "typeVersion": 4.2
    },
    {
      "id": "c2b91608-a4bc-4c65-9697-af3d05e6b097",
      "name": "发送文本消息",
      "type": "n8n-nodes-base.telegram",
      "position": [
        2300,
        20
      ],
      "parameters": {
        "text": "=💳 New transaction : \n- Date : {{ $json.DATE }}\n- Transaction detail : {{ $json['TRANSACTION DETAIL'] }}\n- Vendor : {{ $json.VENDOR }}\n- Total transaction : {{ $json.VALUE }}\n\n---",
        "chatId": "YOUR_TELEGRAM_CHAT_ID",
        "additionalFields": {
          "appendAttribution": false
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "89ae0865-ac70-4cdb-8f14-e64dd023ede2",
      "name": "上传文件",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        520,
        20
      ],
      "parameters": {
        "name": "={{ $json.Image[0].filename }}",
        "driveId": {
          "__rl": true,
          "mode": "list",
          "value": "My Drive"
        },
        "options": {},
        "folderId": {
          "__rl": true,
          "mode": "list",
          "value": "YOUR_GOOGLE_DRIVE_FOLDER_ID"
        },
        "inputDataFieldName": "=Image"
      },
      "typeVersion": 3
    },
    {
      "id": "2dc29bd4-058e-426f-a78c-e99b15f67c82",
      "name": "代码",
      "type": "n8n-nodes-base.code",
      "position": [
        940,
        20
      ],
      "parameters": {
        "mode": "runOnceForEachItem",
        "jsCode": "// Convert file to base64 for Vision API\nconst base64 = item.binary.data.data; // 'data' is the binary name\nreturn {\n  json: {\n    base64\n  }\n};\n"
      },
      "typeVersion": 2
    },
    {
      "id": "a755e54a-5361-46aa-8db0-089a22195b96",
      "name": "代码1",
      "type": "n8n-nodes-base.code",
      "position": [
        1880,
        20
      ],
      "parameters": {
        "jsCode": "function cleanNumber(n) {\n  return parseInt(n.toString().replace(/[.,]/g, '')) || 0;\n}\n\n// Get output from $json\nconst output = $json.output || {};\n\n// If there are Items, process them. If not, skip.\nconst items = Array.isArray(output.Items)\n  ? output.Items.map(item => ({\n      ...item,\n      Quantity: cleanNumber(item.Quantity),\n      \"Unit Price\": cleanNumber(item[\"Unit Price\"]),\n      Total: cleanNumber(item.Total),\n    }))\n  : undefined;\n\n// Optional: clean Total Amount as well\nconst totalAmount = output[\"Total Amount\"]\n  ? cleanNumber(output[\"Total Amount\"])\n  : 0;\n\nreturn {\n  json: {\n    output: {\n      ...output,\n      ...(items && { Items: items }),\n      \"Total Amount\": totalAmount,\n    }\n  }\n};\n"
      },
      "typeVersion": 2
    },
    {
      "id": "606d3086-073c-40b5-b65e-c9de3101d86d",
      "name": "在表格中追加或更新行",
      "type": "n8n-nodes-base.googleSheets",
      "position": [
        2100,
        20
      ],
      "parameters": {
        "columns": {
          "value": {
            "ID": "={{ $('Download file').item.json.id }}",
            "IMG": "={{ $('Download file').item.json.webViewLink }}",
            "DATE": "={{ $json.output.DATE }}",
            "VALUE": "={{ $json.output.VALUE }}",
            "VENDOR": "={{ $json.output.VENDOR }}",
            "FILE NAME": "={{ $('Download file').item.json.name }}",
            "NO VOUCHER": "={{ $json.output['NO VOUCHER'] }}",
            "TRANSACTION DETAIL": "={{ $json.output['TRANSACTION DETAIL'] }}"
          },
          "schema": [
            {
              "id": "ID",
              "type": "string",
              "display": true,
              "removed": false,
              "required": false,
              "displayName": "ID",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "FILE NAME",
              "type": "string",
              "display": true,
              "required": false,
              "displayName": "FILE NAME",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "IMG",
              "type": "string",
              "display": true,
              "required": false,
              "displayName": "IMG",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "DATE",
              "type": "string",
              "display": true,
              "removed": false,
              "required": false,
              "displayName": "DATE",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "NO VOUCHER",
              "type": "string",
              "display": true,
              "required": false,
              "displayName": "NO VOUCHER",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "TRANSACTION DETAIL",
              "type": "string",
              "display": true,
              "removed": false,
              "required": false,
              "displayName": "TRANSACTION DETAIL",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "VENDOR",
              "type": "string",
              "display": true,
              "required": false,
              "displayName": "VENDOR",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            },
            {
              "id": "VALUE",
              "type": "string",
              "display": true,
              "removed": false,
              "required": false,
              "displayName": "VALUE",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            }
          ],
          "mappingMode": "defineBelow",
          "matchingColumns": [
            "ID"
          ],
          "attemptToConvertTypes": false,
          "convertFieldsToString": false
        },
        "options": {},
        "operation": "appendOrUpdate",
        "sheetName": {
          "__rl": true,
          "mode": "list",
          "value": "gid=0"
        },
        "documentId": {
          "__rl": true,
          "mode": "list",
          "value": "YOUR_GOOGLE_SHEETS_DOCUMENT_ID"
        }
      },
      "typeVersion": 4.6
    },
    {
      "id": "71add4c5-64ef-49e2-9fbf-7b6034b17705",
      "name": "表单提交时",
      "type": "n8n-nodes-base.formTrigger",
      "position": [
        300,
        20
      ],
      "parameters": {
        "options": {
          "appendAttribution": false
        },
        "formTitle": "Upload Purchase Invoice",
        "formFields": {
          "values": [
            {
              "fieldType": "file",
              "fieldLabel": "Image",
              "requiredField": true,
              "acceptFileTypes": ".jpg, .png, .jpeg"
            }
          ]
        },
        "responseMode": "lastNode"
      },
      "typeVersion": 2.2
    },
    {
      "id": "98b5b08e-2b7c-4360-a224-43fd6059bb34",
      "name": "便签2",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        260,
        -200
      ],
      "parameters": {
        "color": 2,
        "width": 600,
        "height": 400,
        "content": "- 当用户提交名为\"上传采购发票\"的表单时，此节点触发工作流。"
      },
      "typeVersion": 1
    },
    {
      "id": "4e897daa-cde8-4d25-8217-9f722c68fe4d",
      "name": "便签",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        900,
        -200
      ],
      "parameters": {
        "color": 4,
        "width": 560,
        "height": 400,
        "content": "- 将下载的发票文件（二进制）转换为 Base64 格式，以便发送到 Google Vision API。"
      },
      "typeVersion": 1
    },
    {
      "id": "510e3239-fbbd-4b47-9c31-813994c3e251",
      "name": "便签3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1480,
        -200
      ],
      "parameters": {
        "color": 5,
        "width": 560,
        "height": 580,
        "content": "- 通过 LLM 处理 OCR 文本，提取并结构化关键发票详情。"
      },
      "typeVersion": 1
    },
    {
      "id": "71dcd4e0-84cc-4b20-ac60-6701cc4e5f60",
      "name": "便签4",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        2060,
        -200
      ],
      "parameters": {
        "color": 6,
        "width": 380,
        "height": 580,
        "content": "- 将提取和清理后的发票数据写入 Google Sheets 文档，添加新行或更新现有行。"
      },
      "typeVersion": 1
    }
  ],
  "active": false,
  "pinData": {},
  "settings": {
    "executionOrder": "v1"
  },
  "versionId": "a58b0780-6c63-40bb-84bd-a621579d6eef",
  "connections": {
    "Code": {
      "main": [
        [
          {
            "node": "Set Vision API",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Code1": {
      "main": [
        [
          {
            "node": "Append or update row in sheet",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Upload file": {
      "main": [
        [
          {
            "node": "Download file",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "HTTP Request": {
      "main": [
        [
          {
            "node": "Basic LLM Chain",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Download file": {
      "main": [
        [
          {
            "node": "Code",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Set Vision API": {
      "main": [
        [
          {
            "node": "HTTP Request",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Basic LLM Chain": {
      "main": [
        [
          {
            "node": "Code1",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "On form submission": {
      "main": [
        [
          {
            "node": "Upload file",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "OpenRouter Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "Basic LLM Chain",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Structured Output Parser": {
      "ai_outputParser": [
        [
          {
            "node": "Basic LLM Chain",
            "type": "ai_outputParser",
            "index": 0
          }
        ]
      ]
    },
    "Append or update row in sheet": {
      "main": [
        [
          {
            "node": "Send a text message",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}
常见问题

如何使用这个工作流？

复制上方的 JSON 配置代码，在您的 n8n 实例中创建新工作流并选择「从 JSON 导入」，粘贴配置后根据需要修改凭证设置即可。

这个工作流适合什么场景？

高级 - 发票处理, 多模态 AI

需要付费吗？

本工作流完全免费，您可以直接导入使用。但请注意，工作流中使用的第三方服务（如 OpenAI API）可能需要您自行付费。

相关工作流推荐

YouTube评论情感和关键词提取器
使用Gemini AI分析YouTube评论情感和关键词并通过Telegram报告
Set
Code
Telegram
+10
20 节点Budi SJ
市场调研
使用Google Vision OCR、Telegram和Google Sheets的自动化财务报告
使用Google Vision OCR、AI和Telegram处理收据到Google表格
If
Set
Code
+12
31 节点Budi SJ
发票处理
自动化视频生成器
使用 DeepSeek、TTS 和 Together.ai 从脚本生成 AI 视频
Set
Code
Wait
+15
81 节点Lakindu Siriwardana
内容创作
使用AI和RSS的多平台YouTube内容生成器
使用AI和RSS从YouTube生成多平台内容
If
Set
Code
+10
37 节点Budi SJ
内容创作
基于 AI、社交媒体和 WhatsApp 链接的自动化本地商业线索生成器
使用 AI、社交媒体和 WhatsApp 链接自动生成本地商业线索
If
Code
Limit
+9
29 节点Budi SJ
潜在客户开发
自动化应用分析与ASO报告生成器
使用Gemini AI和Google Docs从Google Play应用生成ASO报告
Code
Telegram
Google Docs
+5
13 节点Budi SJ
市场调研
工作流信息
难度等级
高级
节点数量17
分类2
节点类型11
难度说明

适合高级用户，包含 16+ 个节点的复杂工作流

作者
Budi SJ

Budi SJ

@budisj

I’m a Product Designer who also works as an Automation Developer. With a background in product design and systems thinking, I build user-centered workflows. My focus is on helping teams and businesses work more productively through impactful automation systems.

外部链接
在 n8n.io 查看 →

分享此工作流