Axion/n8n-workflows/receipt-ocr-workflow.json

{
  "name": "Receipt OCR Analysis",
  "nodes": [
    {
      "parameters": {
        "httpMethod": "POST",
        "path": "receipt-upload",
        "responseMode": "responseNode",
        "options": {}
      },
      "id": "webhook-receipt-upload",
      "name": "Webhook - Receipt Upload",
      "type": "n8n-nodes-base.webhook",
      "typeVersion": 1,
      "position": [250, 300],
      "webhookId": "receipt-upload-webhook"
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "extract-image",
              "name": "image",
              "value": "={{ $json.body.image }}",
              "type": "string"
            },
            {
              "id": "extract-user-id",
              "name": "userId",
              "value": "={{ $json.body.userId }}",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "id": "extract-data",
      "name": "Extract Receipt Data",
      "type": "n8n-nodes-base.set",
      "typeVersion": 1,
      "position": [450, 300]
    },
    {
      "parameters": {
        "authentication": "genericCredentialType",
        "genericAuthType": "httpHeaderAuth",
        "httpHeaderAuth": {
          "name": "Authorization",
          "value": "Bearer {{ $env.OCR_API_KEY }}"
        },
        "requestMethod": "POST",
        "url": "https://api.ocr.space/parse/image",
        "sendBody": true,
        "bodyParameters": {
          "parameters": [
            {
              "name": "apikey",
              "value": "={{ $env.OCR_API_KEY }}"
            },
            {
              "name": "base64Image",
              "value": "={{ $json.image }}"
            },
            {
              "name": "language",
              "value": "eng"
            },
            {
              "name": "isOverlayRequired",
              "value": "false"
            }
          ]
        },
        "options": {}
      },
      "id": "ocr-api-call",
      "name": "OCR API Call",
      "type": "n8n-nodes-base.httpRequest",
      "typeVersion": 1,
      "position": [650, 300]
    },
    {
      "parameters": {
        "jsCode": "// Extract text from OCR response\nconst ocrResponse = $input.item.json;\nconst parsedText = ocrResponse.ParsedResults?.[0]?.ParsedText || '';\n\n// Regular expressions to extract receipt data\nconst amountRegex = /(?:total|amount|sum|\\$|€|£)\\s*:?\\s*([\\d,]+\\.[\\d]{2})/i;\nconst dateRegex = /(\\d{1,2}[\\/\\-]\\d{1,2}[\\/\\-]\\d{2,4})/;\nconst vendorRegex = /^([A-Z][A-Za-z\\s&]+?)(?:\\s|$)/m;\nconst taxRegex = /(?:tax|vat|gst)\\s*:?\\s*([\\d,]+\\.[\\d]{2})/i;\n\n// Extract amount\nlet amount = null;\nconst amountMatch = parsedText.match(amountRegex);\nif (amountMatch) {\n  amount = parseFloat(amountMatch[1].replace(/,/g, ''));\n}\n\n// Extract date\nlet date = null;\nconst dateMatch = parsedText.match(dateRegex);\nif (dateMatch) {\n  date = dateMatch[1];\n}\n\n// Extract vendor (first line or company name)\nlet vendor = null;\nconst vendorMatch = parsedText.match(vendorRegex);\nif (vendorMatch) {\n  vendor = vendorMatch[1].trim();\n}\n\n// Extract tax\nlet tax = null;\nconst taxMatch = parsedText.match(taxRegex);\nif (taxMatch) {\n  tax = parseFloat(taxMatch[1].replace(/,/g, ''));\n}\n\n// Calculate confidence score based on extracted data\nlet confidence = 0;\nif (amount) confidence += 0.4;\nif (date) confidence += 0.2;\nif (vendor) confidence += 0.2;\nif (tax) confidence += 0.2;\n\nreturn {\n  json: {\n    userId: $input.item.json.userId,\n    originalImage: $input.item.json.image,\n    extractedText: parsedText,\n    amount: amount,\n    date: date,\n    vendor: vendor,\n    tax: tax || 0,\n    confidence: confidence,\n    status: confidence >= 0.6 ? 'needs_review' : 'pending',\n    ocrRawResponse: ocrResponse\n  }\n};"
      },
      "id": "parse-receipt-data",
      "name": "Parse Receipt Data",
      "type": "n8n-nodes-base.code",
      "typeVersion": 1,
      "position": [850, 300]
    },
    {
      "parameters": {
        "authentication": "genericCredentialType",
        "genericAuthType": "httpHeaderAuth",
        "httpHeaderAuth": {
          "name": "Authorization",
          "value": "Bearer {{ $env.BACKEND_API_KEY }}"
        },
        "requestMethod": "POST",
        "url": "={{ $env.BACKEND_API_URL }}/api/receipts",
        "sendBody": true,
        "bodyParameters": {
          "parameters": [
            {
              "name": "userId",
              "value": "={{ $json.userId }}"
            },
            {
              "name": "amount",
              "value": "={{ $json.amount }}"
            },
            {
              "name": "date",
              "value": "={{ $json.date }}"
            },
            {
              "name": "vendor",
              "value": "={{ $json.vendor }}"
            },
            {
              "name": "tax",
              "value": "={{ $json.tax }}"
            },
            {
              "name": "confidence",
              "value": "={{ $json.confidence }}"
            },
            {
              "name": "status",
              "value": "={{ $json.status }}"
            },
            {
              "name": "extractedText",
              "value": "={{ $json.extractedText }}"
            }
          ]
        },
        "options": {}
      },
      "id": "save-to-backend",
      "name": "Save to Backend",
      "type": "n8n-nodes-base.httpRequest",
      "typeVersion": 1,
      "position": [1050, 300]
    },
    {
      "parameters": {
        "respondWith": "json",
        "responseBody": "={{ { \"success\": true, \"receiptId\": $json.id, \"amount\": $json.amount, \"confidence\": $json.confidence, \"status\": $json.status } }}"
      },
      "id": "respond-success",
      "name": "Respond Success",
      "type": "n8n-nodes-base.respondToWebhook",
      "typeVersion": 1,
      "position": [1250, 300]
    }
  ],
  "connections": {
    "Webhook - Receipt Upload": {
      "main": [
        [
          {
            "node": "Extract Receipt Data",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Extract Receipt Data": {
      "main": [
        [
          {
            "node": "OCR API Call",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "OCR API Call": {
      "main": [
        [
          {
            "node": "Parse Receipt Data",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Parse Receipt Data": {
      "main": [
        [
          {
            "node": "Save to Backend",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Save to Backend": {
      "main": [
        [
          {
            "node": "Respond Success",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  },
  "pinData": {},
  "settings": {
    "executionOrder": "v1"
  },
  "staticData": null,
  "tags": [],
  "triggerCount": 1,
  "updatedAt": "2024-01-01T00:00:00.000Z",
  "versionId": "1"
}