{ "name": "Invoice Extractor", "nodes": [ { "parameters": { "httpMethod": "POST", "path": "invoice-upload", "responseMode": "responseNode", "options": { "rawBody": false } }, "id": "62d46998-3a69-4bd0-b80d-3edf9f54811c", "name": "Webhook Trigger", "type": "n8n-nodes-base.webhook", "typeVersion": 2, "position": [-1472, 80] }, { "parameters": { "operation": "pdf", "binaryPropertyName": "invoice", "options": {} }, "id": "02b6eb9e-f056-48c6-8cfd-91b9064f73da", "name": "Extract Text from PDF", "type": "n8n-nodes-base.extractFromFile", "typeVersion": 1.1, "position": [-1248, 80] }, { "parameters": { "assignments": { "assignments": [ { "id": "713a0531-31a6-402d-ab83-9fb9fb94e21d", "name": "systemPrompt", "type": "string", "value": "Extract the invoice from the provided PDF text.\n\nReturn ONLY valid JSON. No markdown. No explanation.\n\n{\n \"supplier_name\": null,\n \"invoice_number\": null,\n \"invoice_date\": null,\n \"due_date\": null,\n \"subtotal\": null,\n \"tax_amount\": null,\n \"total_amount\": null,\n \"line_items\": [\n {\n \"description\": null,\n \"quantity\": null,\n \"unit_price\": null,\n \"amount\": null\n }\n ]\n}\n\nRules:\n\n- Extract ONLY information explicitly present in the invoice.\n- NEVER guess or invent missing values.\n- If a field is missing, return null.\n- invoice_date and due_date are separate fields.\n- Extract the actual Due Date when one is shown.\n- Extract every line item shown in the invoice.\n- Extract the quantity, unit price, and amount for every line item.\n- Extract the printed subtotal, tax amount, and total exactly as shown.\n- Do NOT calculate or repair values.\n- If the invoice contains inconsistent numbers, preserve the numbers shown and let the validation step detect the inconsistency.\n- Return JSON only." }, { "id": "5724aac1-2b0c-4be1-b40c-b6266095c0ae", "name": "text", "type": "string", "value": "={{ $json.text }}" } ] }, "options": {} }, "id": "4a431769-fcc2-463a-8623-c465987417a8", "name": "Build Prompt", "type": "n8n-nodes-base.set", "typeVersion": 3.4, "position": [-1024, 80] }, { "parameters": { "jsCode": "const item = $input.item.json;\n\nlet rawContent = item.text || '';\n\nlet cleanedText = String(rawContent)\n .replace(/```json/gi, '')\n .replace(/```/g, '')\n .trim();\n\nconst jsonMatch = cleanedText.match(/\\{[\\s\\S]*\\}/);\n\nif (jsonMatch) cleanedText = jsonMatch[0];\n\nfunction toNumber(value) {\n if (value === null || value === undefined || value === '') return null;\n if (typeof value === 'number') return Number.isFinite(value) ? value : null;\n const cleaned = String(value).replace(/[^0-9.-]/g, '');\n if (!cleaned || cleaned === '-') return null;\n const number = parseFloat(cleaned);\n return Number.isFinite(number) ? number : null;\n}\n\nfunction cleanString(value) {\n if (value === null || value === undefined) return null;\n const valueString = String(value).trim();\n return valueString === '' ? null : valueString;\n}\n\nfunction matches(a, b) {\n if (a === null || b === null) return false;\n return Math.abs(a - b) <= 0.02;\n}\n\nlet parsed = {};\nlet parseError = null;\ntry { parsed = JSON.parse(cleanedText); } catch (e) { parseError = e.message; }\n\nconst cleaned = {\n supplier_name: cleanString(parsed.supplier_name),\n invoice_number: cleanString(parsed.invoice_number),\n invoice_date: cleanString(parsed.invoice_date),\n due_date: cleanString(parsed.due_date),\n subtotal: toNumber(parsed.subtotal),\n tax_amount: toNumber(parsed.tax_amount),\n total_amount: toNumber(parsed.total_amount),\n line_items: Array.isArray(parsed.line_items) ? parsed.line_items.map(line => ({\n description: cleanString(line.description),\n quantity: toNumber(line.quantity),\n unit_price: toNumber(line.unit_price),\n amount: toNumber(line.amount)\n })) : []\n};\n\nconst errors = [];\nif (parseError) errors.push('Invalid JSON returned by AI');\nif (!cleaned.supplier_name) errors.push('Missing supplier name');\nif (!cleaned.invoice_number) errors.push('Missing invoice number');\nif (!cleaned.invoice_date) errors.push('Missing invoice date');\nif (!cleaned.due_date) errors.push('Missing due date');\nif (cleaned.subtotal === null) errors.push('Missing subtotal');\nif (cleaned.tax_amount === null) errors.push('Missing tax amount');\nif (cleaned.total_amount === null) errors.push('Missing total amount');\n\nif (!cleaned.line_items.length) {\n errors.push('No line items found');\n} else {\n cleaned.line_items.forEach((line, index) => {\n const n = index + 1;\n if (!line.description) errors.push(`Line ${n}: missing description`);\n if (line.quantity === null) errors.push(`Line ${n}: missing quantity`);\n if (line.unit_price === null) errors.push(`Line ${n}: missing unit price`);\n if (line.amount === null) errors.push(`Line ${n}: missing amount`);\n if (line.quantity !== null && line.unit_price !== null && line.amount !== null) {\n const expected = line.quantity * line.unit_price;\n if (!matches(expected, line.amount)) errors.push(`Line ${n}: quantity × unit price does not equal amount`);\n }\n });\n}\n\nif (cleaned.line_items.length > 0 && cleaned.line_items.every(line => line.amount !== null) && cleaned.subtotal !== null) {\n const calculatedSubtotal = cleaned.line_items.reduce((sum, line) => sum + line.amount, 0);\n if (!matches(calculatedSubtotal, cleaned.subtotal)) errors.push('Subtotal does not match line items');\n}\n\nif (cleaned.subtotal !== null && cleaned.tax_amount !== null && cleaned.total_amount !== null) {\n const calculatedTotal = cleaned.subtotal + cleaned.tax_amount;\n if (!matches(calculatedTotal, cleaned.total_amount)) errors.push('Subtotal + tax does not equal total');\n}\n\nlet sourceFilename = 'unknown';\ntry {\n const binary = $('Webhook Trigger').first().binary;\n const key = Object.keys(binary || {})[0];\n if (key && binary[key].fileName) sourceFilename = binary[key].fileName;\n} catch (e) {}\n\nreturn [{ json: { ...cleaned, source_filename: sourceFilename, processed_at: new Date().toISOString(), valid: !parseError && errors.length === 0, error_reason: errors.length ? errors.join('; ') : null, validation_errors: errors, raw_ai_response: rawContent } }];" }, "id": "5b6941db-a1f4-4d42-98ff-0b087b4a68db", "name": "Clean & Validate Data", "type": "n8n-nodes-base.code", "typeVersion": 2, "position": [-448, 80] }, { "parameters": { "conditions": { "options": {"caseSensitive": true,"leftValue": "","typeValidation": "strict","version": 2}, "conditions": [{"id": "b674a597-7eae-449b-8d91-731c886e6ff1","leftValue": "={{ $json.valid }}","rightValue": true,"operator": {"type": "boolean","operation": "true","singleValue": true}}], "combinator": "and" }, "options": {} }, "id": "8e42e6f0-86f0-4d1e-8d0c-0f95e6779f13", "name": "Is Data Valid?", "type": "n8n-nodes-base.if", "typeVersion": 2.2, "position": [-224, 80] }, { "parameters": { "operation": "append", "documentId": {"__rl": true,"value": "1zemmx7Hd66hFTFRIaLshxynjhkRyawaPZoKq3sg5naU","mode": "list","cachedResultName": "Invoice Logged","cachedResultUrl": "https://docs.google.com/spreadsheets/d/YOUR_GOOGLE_SHEET_ID/edit"}, "sheetName": {"__rl": true,"value": "gid=0","mode": "list","cachedResultName": "Sheet1","cachedResultUrl": "https://docs.google.com/spreadsheets/d/YOUR_GOOGLE_SHEET_ID/edit"}, "columns": {"mappingMode": "defineBelow","value": {"Invoice Number": "={{ $json.invoice_number }}","Vendor Name": "={{ $json.supplier_name }}","Total Amount": "={{ $json.total_amount }}","Tax Amount": "={{ $json.tax_amount }}","Due Date": "={{ $json.due_date }}","Validation Status": "={{ $json.valid }}","Invoice Date": "={{$json.invoice_date}}"},"matchingColumns": [],"schema": [],"attemptToConvertTypes": false,"convertFieldsToString": false}, "options": {} }, "id": "2dce12ef-53a3-486a-9830-f263d9d817a9", "name": "Append to Google Sheet", "type": "n8n-nodes-base.googleSheets", "typeVersion": 4.5, "position": [0, -16] }, { "parameters": {"respondWith": "json","responseBody": "={{ { \"success\": true, \"message\": \"Invoice processed and logged.\", \"data\": $json } }}","options": {"responseCode": 200}}, "id": "cb1346a4-5e6a-4a58-b4e3-34fa3739a875", "name": "Respond - Success", "type": "n8n-nodes-base.respondToWebhook", "typeVersion": 1.1, "position": [224, -16] }, { "parameters": {"respondWith": "json","responseBody": "={{ { \"success\": false, \"message\": \"Could not confidently extract all required fields. Please review manually.\", \"reason\": $json.error_reason, \"raw_ai_response\": $json.raw_ai_response } }}","options": {"responseCode": 422}}, "id": "20d113a9-cac0-4be3-b4ca-c38abd389643", "name": "Respond - Needs Review", "type": "n8n-nodes-base.respondToWebhook", "typeVersion": 1.1, "position": [0, 176] }, { "parameters": {"content": "## Invoice Extraction Pipeline\n\n**Before running:**\n1. Configure your OpenRouter credential on the AI node.\n2. Configure your Google Sheets credential, spreadsheet ID, and sheet/tab name.\n3. Use the Webhook Production URL from the Webhook Trigger node.\n4. This pipeline only reads text-based PDFs. Scanned/photographed invoices need a vision-capable model step instead.","height": 300,"width": 460}, "id": "c847eee4-b063-45a8-acba-7c35f497de33", "name": "Sticky Note", "type": "n8n-nodes-base.stickyNote", "typeVersion": 1, "position": [-1456, 384] }, { "parameters": {"promptType": "define","text": "=You are a strict invoice data extraction tool.\n\nRead the invoice text below and extract these fields:\n\n- supplier_name\n- invoice_number\n- invoice_date\n- due_date\n- subtotal\n- tax_amount\n- total_amount\n- line_items\n\nRules:\n1. Return ONLY valid JSON.\n2. Do not use markdown.\n3. Do not add explanations.\n4. If a field cannot be found, return null.\n5. line_items must be an array. If no line items can be found, return [].\n6. Do not invent or calculate missing invoice information.\n7. Preserve the values exactly as shown when possible.\n8. For monetary values, return numbers without currency symbols or commas.\n\nInvoice text:\n\n{{ $json.text }}","batching": {}}, "type": "@n8n/n8n-nodes-langchain.chainLlm", "typeVersion": 1.9, "position": [-800, 80], "id": "44b51b58-fd5c-40ca-bb8d-7509296db6ab", "name": "Basic LLM Chain" }, { "parameters": {"model": "nvidia/nemotron-3-nano-30b-a3b:free","options": {}}, "type": "@n8n/n8n-nodes-langchain.lmChatOpenRouter", "typeVersion": 1, "position": [-736, 320], "id": "d02bda24-d001-44f3-91cb-f7f5544d0988", "name": "OpenRouter Chat Model" } ], "connections": { "Webhook Trigger": {"main": [[{"node": "Extract Text from PDF","type": "main","index": 0}]]}, "Extract Text from PDF": {"main": [[{"node": "Build Prompt","type": "main","index": 0}]]}, "Build Prompt": {"main": [[{"node": "Basic LLM Chain","type": "main","index": 0}]]}, "Clean & Validate Data": {"main": [[{"node": "Is Data Valid?","type": "main","index": 0}]]}, "Is Data Valid?": {"main": [[{"node": "Append to Google Sheet","type": "main","index": 0}],[{"node": "Respond - Needs Review","type": "main","index": 0}]]}, "Append to Google Sheet": {"main": [[{"node": "Respond - Success","type": "main","index": 0}]]}, "OpenRouter Chat Model": {"ai_languageModel": [[{"node": "Basic LLM Chain","type": "ai_languageModel","index": 0}]]}, "Basic LLM Chain": {"main": [[{"node": "Clean & Validate Data","type": "main","index": 0}]]} }, "settings": {"executionOrder": "v1","binaryMode": "separate","availableInMCP": true,"timeSavedMode": "fixed"}, "nodeGroups": [], "tags": [] }