{ "name": "AI Document Processor - Intelligent Data Extraction", "nodes": [ { "parameters": { "httpMethod": "POST", "path": "document-upload", "responseMode": "onReceived", "options": { "rawBody": true } }, "id": "webhook-upload", "name": "Document Upload Webhook", "type": "n8n-nodes-base.webhook", "typeVersion": 2, "position": [ 250, 400 ] }, { "parameters": { "jsCode": "// Process incoming document upload\nconst data = $input.item.json;\nconst body = data.body || {};\n\n// Extract file info from webhook\nconst fileName = body.fileName || data.fileName || 'unknown.pdf';\nconst fileUrl = body.fileUrl || data.fileUrl || '';\nconst fileType = fileName.split('.').pop().toLowerCase();\nconst uploadedBy = body.uploadedBy || data.uploadedBy || 'Unknown';\nconst documentType = body.documentType || 'unknown'; // invoice, contract, brief, etc.\n\n// Determine processing strategy based on file type\nlet processingMethod = 'ocr';\nif (['pdf', 'docx', 'doc'].includes(fileType)) {\n processingMethod = 'text_extraction';\n} else if (['jpg', 'jpeg', 'png', 'gif'].includes(fileType)) {\n processingMethod = 'ocr';\n}\n\nreturn {\n processing_id: `DOC-${Date.now()}`,\n file_name: fileName,\n file_url: fileUrl,\n file_type: fileType,\n document_type: documentType,\n uploaded_by: uploadedBy,\n processing_method: processingMethod,\n uploaded_at: new Date().toISOString(),\n status: 'processing'\n};" }, "id": "parse-upload", "name": "Parse Document Metadata", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ 470, 400 ] }, { "parameters": { "url": "={{ $json.file_url }}", "options": { "response": { "response": { "fullResponse": false, "responseFormat": "file" } } } }, "id": "download-file", "name": "Download Document", "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ 690, 400 ] }, { "parameters": { "operation": "extractText", "options": {} }, "id": "extract-text", "name": "Extract Text from PDF", "type": "n8n-nodes-base.readPDF", "typeVersion": 1, "position": [ 910, 300 ] }, { "parameters": { "url": "https://api.ocr.space/parse/image", "sendBody": true, "bodyParameters": { "parameters": [ { "name": "base64Image", "value": "=data:{{ $('Download Document').item.binary.data.mimeType }};base64,{{ $('Download Document').item.binary.data.data }}" }, { "name": "language", "value": "eng" } ] }, "options": {} }, "id": "ocr-scan", "name": "OCR Scan (Images)", "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ 910, 500 ] }, { "parameters": { "mode": "combine", "combinationMode": "mergeByPosition", "options": {} }, "id": "merge-text", "name": "Merge Extracted Text", "type": "n8n-nodes-base.merge", "typeVersion": 3, "position": [ 1130, 400 ] }, { "parameters": { "jsCode": "// Prepare text for AI analysis\nconst metadata = $('Parse Document Metadata').item.json;\nconst extractedText = $input.item.json.text || \n $input.item.json.ParsedResults?.[0]?.ParsedText || \n 'No text extracted';\n\nconst documentType = metadata.document_type;\n\n// Define extraction prompts based on document type\nconst prompts = {\n invoice: `Extract these fields from this invoice:\\n- Invoice Number\\n- Invoice Date\\n- Due Date\\n- Vendor Name\\n- Vendor Address\\n- Total Amount\\n- Tax Amount\\n- Line Items (description, quantity, unit price, total)\\n- Payment Terms\\n- PO Number (if present)`,\n \n contract: `Extract these fields from this contract:\\n- Contract Title\\n- Contract Date\\n- Effective Date\\n- Expiration Date\\n- Party A (name, address)\\n- Party B (name, address)\\n- Contract Value\\n- Payment Terms\\n- Key Obligations\\n- Termination Clauses`,\n \n brief: `Extract these fields from this brief:\\n- Project Name\\n- Client Name\\n- Project Type\\n- Budget\\n- Timeline/Deadline\\n- Objectives\\n- Deliverables\\n- Key Requirements\\n- Contact Person`,\n \n receipt: `Extract these fields from this receipt:\\n- Merchant Name\\n- Transaction Date\\n- Total Amount\\n- Tax Amount\\n- Payment Method\\n- Items Purchased\\n- Receipt Number`,\n \n unknown: `Analyze this document and extract all key structured data including:\\n- Document type\\n- Date(s)\\n- Names/Entities\\n- Amounts/Numbers\\n- Key terms or clauses`\n};\n\nconst prompt = prompts[documentType] || prompts.unknown;\n\nreturn {\n processing_id: metadata.processing_id,\n document_type: documentType,\n extracted_text: extractedText.substring(0, 4000), // Limit for API\n ai_prompt: prompt,\n file_name: metadata.file_name,\n uploaded_by: metadata.uploaded_by\n};" }, "id": "prepare-ai-prompt", "name": "Prepare AI Extraction Prompt", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ 1350, 400 ] }, { "parameters": { "url": "https://api.anthropic.com/v1/messages", "authentication": "genericCredentialType", "genericAuthType": "httpHeaderAuth", "sendHeaders": true, "headerParameters": { "parameters": [ { "name": "anthropic-version", "value": "2023-06-01" }, { "name": "content-type", "value": "application/json" } ] }, "sendBody": true, "contentType": "json", "body": "={\n \"model\": \"claude-3-5-sonnet-20241022\",\n \"max_tokens\": 2048,\n \"messages\": [\n {\n \"role\": \"user\",\n \"content\": \"{{ $json.ai_prompt }}\\n\\nDocument text:\\n{{ $json.extracted_text }}\\n\\nRespond ONLY with valid JSON containing the extracted fields. Use null for missing fields. Ensure all amounts are numbers, all dates are in YYYY-MM-DD format.\"\n }\n ]\n}", "options": {} }, "id": "ai-extraction", "name": "AI Data Extraction (Claude)", "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ 1570, 400 ] }, { "parameters": { "jsCode": "// Parse AI response and structure data\nconst aiResponse = $input.item.json;\nconst metadata = $('Prepare AI Extraction Prompt').item.json;\n\nlet extractedData = {};\n\ntry {\n // Extract JSON from AI response\n const aiText = aiResponse.content?.[0]?.text || '{}';\n \n // Try to parse JSON (handle markdown code blocks if present)\n const jsonMatch = aiText.match(/```json\\n([\\s\\S]*?)\\n```/) || \n aiText.match(/```\\n([\\s\\S]*?)\\n```/) ||\n [null, aiText];\n \n extractedData = JSON.parse(jsonMatch[1]);\n} catch (error) {\n extractedData = {\n error: 'Failed to parse AI response',\n raw_response: aiResponse.content?.[0]?.text?.substring(0, 500)\n };\n}\n\n// Add metadata\nextractedData.processing_id = metadata.processing_id;\nextractedData.file_name = metadata.file_name;\nextractedData.document_type = metadata.document_type;\nextractedData.uploaded_by = metadata.uploaded_by;\nextractedData.processed_at = new Date().toISOString();\nextractedData.ai_model = 'claude-3-5-sonnet';\n\n// Count missing fields\nconst requiredFields = Object.keys(extractedData).filter(k => !['processing_id', 'file_name', 'document_type', 'uploaded_by', 'processed_at', 'ai_model', 'error'].includes(k));\nconst missingFields = requiredFields.filter(field => !extractedData[field] || extractedData[field] === null || extractedData[field] === '');\n\nextractedData.missing_fields_count = missingFields.length;\nextractedData.missing_fields = missingFields.join(', ');\nextractedData.extraction_confidence = missingFields.length === 0 ? 'high' : missingFields.length <= 2 ? 'medium' : 'low';\n\nreturn extractedData;" }, "id": "parse-ai-response", "name": "Parse & Structure Extracted Data", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ 1790, 400 ] }, { "parameters": { "authentication": "serviceAccount", "operation": "appendOrUpdate", "documentId": { "__rl": true, "mode": "list", "value": "" }, "sheetName": { "__rl": true, "mode": "list", "value": "" }, "columns": { "mappingMode": "autoMapInputData", "value": {} }, "options": {} }, "id": "save-to-sheets", "name": "Save to Google Sheets", "type": "n8n-nodes-base.googleSheets", "typeVersion": 4.4, "position": [ 2010, 400 ] }, { "parameters": { "conditions": { "options": { "caseSensitive": true, "leftValue": "", "typeValidation": "strict" }, "conditions": [ { "id": "missing-check", "leftValue": "={{ $json.missing_fields_count }}", "rightValue": 0, "operator": { "type": "number", "operation": "gt" } }, { "id": "confidence-check", "leftValue": "={{ $json.extraction_confidence }}", "rightValue": "low", "operator": { "type": "string", "operation": "equals", "singleValue": true } } ], "combinator": "or" }, "options": {} }, "id": "check-completeness", "name": "Check Data Completeness", "type": "n8n-nodes-base.if", "typeVersion": 2, "position": [ 2230, 400 ] }, { "parameters": { "select": "channel", "channelId": { "__rl": true, "mode": "list", "value": "" }, "text": "=\u26a0\ufe0f **Document Review Needed**\\n\\n*Processing ID:* {{ $json.processing_id }}\\n*File:* {{ $json.file_name }}\\n*Type:* {{ $json.document_type }}\\n*Uploaded by:* {{ $json.uploaded_by }}\\n\\n*Issue:* {{ $json.missing_fields_count }} fields missing or incomplete\\n*Missing fields:* {{ $json.missing_fields }}\\n*Confidence:* {{ $json.extraction_confidence }}\\n\\n*Action needed:* Please review and complete the data in Google Sheets.\\n\\n[View in Sheets](SHEET_URL_HERE)", "otherOptions": {} }, "id": "alert-slack-incomplete", "name": "Alert - Incomplete Data (Slack)", "type": "n8n-nodes-base.slack", "typeVersion": 2.2, "position": [ 2450, 300 ] }, { "parameters": { "fromEmail": "documents@company.com", "toEmail": "={{ $json.uploaded_by }}", "subject": "\u26a0\ufe0f Document Processed - Review Required: {{ $json.file_name }}", "text": "=Hi,\\n\\nYour document has been processed, but we need your help to complete some information.\\n\\n**Document:** {{ $json.file_name }}\\n**Processing ID:** {{ $json.processing_id }}\\n**Type:** {{ $json.document_type }}\\n\\n**Status:** Review Required\\n**Missing fields:** {{ $json.missing_fields }}\\n\\nWe've extracted as much data as possible, but {{ $json.missing_fields_count }} field(s) need your attention.\\n\\nPlease review the data in our system and fill in the missing information.\\n\\nIf you have questions, reply to this email or contact support.\\n\\nThanks,\\nAutomated Document Processing Team", "options": { "senderName": "Document Processor" } }, "id": "email-review-needed", "name": "Email - Review Needed", "type": "n8n-nodes-base.sendGrid", "typeVersion": 1, "position": [ 2450, 500 ] }, { "parameters": { "fromEmail": "documents@company.com", "toEmail": "={{ $json.uploaded_by }}", "subject": "\u2705 Document Successfully Processed: {{ $json.file_name }}", "emailFormat": "html", "html": "=\\n\\n\\n \\n\\n\\n
\\n

\u2705 Document Successfully Processed

\\n
\\n
\\n

Hi,

\\n

Your document has been successfully processed and all data extracted!

\\n \\n
\\n
Document:
\\n {{ $json.file_name }}\\n
\\n \\n
\\n
Processing ID:
\\n {{ $json.processing_id }}\\n
\\n \\n
\\n
Document Type:
\\n {{ $json.document_type }}\\n
\\n \\n
\\n
Processed:
\\n {{ new Date($json.processed_at).toLocaleString() }}\\n
\\n \\n
\\n
Confidence:
\\n {{ $json.extraction_confidence }} ({{ $json.missing_fields_count === 0 ? 'All fields extracted' : 'Some fields incomplete' }})\\n
\\n \\n

The extracted data has been saved to your database and is ready for use.

\\n \\n

If you notice any errors, please contact support with your Processing ID.

\\n \\n

Thanks,
Automated Document Processing Team

\\n
\\n\\n", "options": { "senderName": "Document Processor" } }, "id": "email-success", "name": "Email - Success Confirmation", "type": "n8n-nodes-base.sendGrid", "typeVersion": 1, "position": [ 2450, 100 ] }, { "parameters": { "select": "channel", "channelId": { "__rl": true, "mode": "list", "value": "" }, "text": "=\u2705 **Document Processed Successfully**\\n\\n*File:* {{ $json.file_name }}\\n*Type:* {{ $json.document_type }}\\n*Uploaded by:* {{ $json.uploaded_by }}\\n*Confidence:* {{ $json.extraction_confidence }}\\n\\nAll fields extracted and saved to Google Sheets.", "otherOptions": {} }, "id": "notify-slack-success", "name": "Notify - Success (Slack)", "type": "n8n-nodes-base.slack", "typeVersion": 2.2, "position": [ 2450, 200 ] }, { "parameters": { "operation": "appendOrUpdate", "documentId": { "__rl": true, "mode": "list", "value": "" }, "sheetName": { "__rl": true, "mode": "list", "value": "" }, "columns": { "mappingMode": "defineBelow", "value": { "Timestamp": "={{ $now.toISO() }}", "Processing ID": "={{ $json.processing_id }}", "File Name": "={{ $json.file_name }}", "Document Type": "={{ $json.document_type }}", "Uploaded By": "={{ $json.uploaded_by }}", "Status": "={{ $json.missing_fields_count === 0 ? 'Complete' : 'Incomplete' }}", "Missing Fields": "={{ $json.missing_fields || 'None' }}", "Confidence": "={{ $json.extraction_confidence }}" } }, "options": {} }, "id": "log-processing", "name": "Log to Audit Trail", "type": "n8n-nodes-base.googleSheets", "typeVersion": 4.4, "position": [ 2670, 400 ] }, { "parameters": {}, "id": "error-trigger", "name": "Error Trigger", "type": "n8n-nodes-base.errorTrigger", "typeVersion": 1, "position": [ 250, 700 ] }, { "parameters": { "jsCode": "// Format error details for notification\nconst error = $input.item.json.error || {};\nconst execution = $input.item.json.execution || {};\n\nreturn {\n error_type: error.name || 'Unknown Error',\n error_message: error.message || 'No error message available',\n error_stack: error.stack?.substring(0, 500) || '',\n failed_node: error.node?.name || 'Unknown Node',\n workflow_id: execution.id || 'unknown',\n workflow_name: 'AI Document Processor',\n timestamp: new Date().toISOString(),\n severity: 'high'\n};" }, "id": "format-error", "name": "Format Error Details", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [ 470, 700 ] }, { "parameters": { "select": "channel", "channelId": { "__rl": true, "mode": "list", "value": "" }, "text": "=\ud83d\udea8 **Workflow Error Alert**\\n\\n*Workflow:* {{ $json.workflow_name }}\\n*Failed Node:* {{ $json.failed_node }}\\n*Error Type:* {{ $json.error_type }}\\n*Time:* {{ $json.timestamp }}\\n\\n*Error Message:*\\n```\\n{{ $json.error_message }}\\n```\\n\\n*Action needed:* Check workflow execution logs for details.\\n*Execution ID:* {{ $json.workflow_id }}", "otherOptions": {} }, "id": "alert-error-slack", "name": "Alert - Error (Slack)", "type": "n8n-nodes-base.slack", "typeVersion": 2.2, "position": [ 690, 700 ] }, { "parameters": { "fromEmail": "alerts@company.com", "toEmail": "devops@company.com", "subject": "\ud83d\udea8 Document Processor Workflow Error", "text": "=Workflow Error Alert\\n\\nWorkflow: {{ $json.workflow_name }}\\nFailed Node: {{ $json.failed_node }}\\nError Type: {{ $json.error_type }}\\nTime: {{ $json.timestamp }}\\n\\nError Message:\\n{{ $json.error_message }}\\n\\nExecution ID: {{ $json.workflow_id }}\\n\\nPlease investigate and resolve.", "options": { "senderName": "n8n Alerts" } }, "id": "email-error-alert", "name": "Email - Error Alert", "type": "n8n-nodes-base.sendGrid", "typeVersion": 1, "position": [ 690, 850 ] } ], "connections": { "Document Upload Webhook": { "main": [ [ { "node": "Parse Document Metadata", "type": "main", "index": 0 } ] ] }, "Parse Document Metadata": { "main": [ [ { "node": "Download Document", "type": "main", "index": 0 } ] ] }, "Download Document": { "main": [ [ { "node": "Extract Text from PDF", "type": "main", "index": 0 }, { "node": "OCR Scan (Images)", "type": "main", "index": 0 } ] ] }, "Extract Text from PDF": { "main": [ [ { "node": "Merge Extracted Text", "type": "main", "index": 0 } ] ] }, "OCR Scan (Images)": { "main": [ [ { "node": "Merge Extracted Text", "type": "main", "index": 1 } ] ] }, "Merge Extracted Text": { "main": [ [ { "node": "Prepare AI Extraction Prompt", "type": "main", "index": 0 } ] ] }, "Prepare AI Extraction Prompt": { "main": [ [ { "node": "AI Data Extraction (Claude)", "type": "main", "index": 0 } ] ] }, "AI Data Extraction (Claude)": { "main": [ [ { "node": "Parse & Structure Extracted Data", "type": "main", "index": 0 } ] ] }, "Parse & Structure Extracted Data": { "main": [ [ { "node": "Save to Google Sheets", "type": "main", "index": 0 } ] ] }, "Save to Google Sheets": { "main": [ [ { "node": "Check Data Completeness", "type": "main", "index": 0 } ] ] }, "Check Data Completeness": { "main": [ [ { "node": "Email - Success Confirmation", "type": "main", "index": 0 }, { "node": "Notify - Success (Slack)", "type": "main", "index": 0 } ], [ { "node": "Alert - Incomplete Data (Slack)", "type": "main", "index": 0 }, { "node": "Email - Review Needed", "type": "main", "index": 0 } ] ] }, "Email - Success Confirmation": { "main": [ [ { "node": "Log to Audit Trail", "type": "main", "index": 0 } ] ] }, "Notify - Success (Slack)": { "main": [ [ { "node": "Log to Audit Trail", "type": "main", "index": 0 } ] ] }, "Alert - Incomplete Data (Slack)": { "main": [ [ { "node": "Log to Audit Trail", "type": "main", "index": 0 } ] ] }, "Email - Review Needed": { "main": [ [ { "node": "Log to Audit Trail", "type": "main", "index": 0 } ] ] }, "Error Trigger": { "main": [ [ { "node": "Format Error Details", "type": "main", "index": 0 } ] ] }, "Format Error Details": { "main": [ [ { "node": "Alert - Error (Slack)", "type": "main", "index": 0 }, { "node": "Email - Error Alert", "type": "main", "index": 0 } ] ] } }, "settings": { "executionOrder": "v1", "errorWorkflow": "" }, "staticData": null, "tags": [], "triggerCount": 2, "versionId": null }