{ "id": "OgbbC37jdxoezN2E", "name": "Template Creator", "tags": [ { "id": "vHVvr905dk6aqmKq", "name": "Template", "createdAt": "2025-05-16T19:05:27.035Z", "updatedAt": "2025-05-16T19:05:27.035Z" }, { "id": "v0qdZQrIENz1TXgD", "name": "PDF", "createdAt": "2025-08-09T22:56:54.358Z", "updatedAt": "2025-08-09T22:56:54.358Z" }, { "id": "DzVAQjk8ZZxDPLam", "name": "Word", "createdAt": "2026-03-22T19:08:05.375Z", "updatedAt": "2026-03-22T19:08:05.375Z" }, { "id": "refSV3bBr5Plhm0A", "name": "Drafting", "createdAt": "2026-03-22T19:08:34.897Z", "updatedAt": "2026-03-22T19:08:34.897Z" } ], "nodes": [ { "id": "979b6ee5-34b0-4f0a-be3d-d03b55d83df6", "name": "Sticky \u2014 Overview", "type": "n8n-nodes-base.stickyNote", "position": [ 0, -32 ], "parameters": { "width": 500, "height": 816, "content": "## \ud83d\udcc4Template Creator\n\n### How it works\nThis workflow accepts any uploaded document (PDF or DOCX) via webhook and automatically converts it into a reusable fill-in-the-blank template.\n\n**Step 1 \u2014 Identify:** GPT-4o first reads the document and determines the document type (e.g., Employment Contract, Invoice, NDA, Lease Agreement, Project Proposal) and the specific variable fields that type of document typically contains.\n\n**Step 2 \u2014 Templatize:** A second AI pass uses the identified document type and field list to replace all variable content with clearly labeled `[BRACKET]` placeholders while preserving all static boilerplate and structure verbatim.\n\n**Step 3 \u2014 Deliver:** The cleaned template is rendered to PDF via Gotenberg, uploaded to Google Drive, made publicly accessible, and a JSON response with file URLs is returned to the caller.\n\n### Setup\n1. Configure the **Webhook** path if needed (default: `general-template-creator`)\n2. Set your **OpenAI** credential on both LLM nodes\n3. Set your **Google Drive** credential and confirm the target folder ID in the Upload node\n4. Confirm the **Gotenberg** URL matches your self-hosted instance\n5. Install the community node `n8n-nodes-word2text` (see \u26a0\ufe0f warning sticky)\n\n### Customization\n- Swap GPT-4o for GPT-4.1 or GPT-4.1-mini on the Identify node to reduce cost on the lighter classification task\n- Add a Switch node after identification to route different document types to type-specific prompts\n- Modify the Drive folder ID to sort templates into subfolders by document type\n\nDocument accepts input from a form such as the one found here:\n[Sample Form](https://iportgpt.com/n8n_assets/template_creator_form.html)\n\n" }, "typeVersion": 1 }, { "id": "d8eea0b3-cc82-4b8b-b557-20ae9d420aaa", "name": "Sticky \u2014 Community Node Warning", "type": "n8n-nodes-base.stickyNote", "position": [ 800, 496 ], "parameters": { "color": 3, "width": 290, "height": 238, "content": "## \u26a0\ufe0f Community Node Required\n`n8n-nodes-word2text` must be installed on your self-hosted n8n instance.\n\n**Settings \u2192 Community Nodes \u2192 Install**\nPackage name: `n8n-nodes-word2text`\n\nHandles the .docx \u2192 plain text extraction path." }, "typeVersion": 1 }, { "id": "987863d3-5ef2-4bce-86bc-4593b7a10186", "name": "Sticky \u2014 Intake", "type": "n8n-nodes-base.stickyNote", "position": [ 512, -32 ], "parameters": { "color": 7, "width": 224, "height": 502, "content": "## 1 \u00b7 Intake\nReceive document upload via POST webhook.\nExpects `fileName` and a binary `file` attachment.\nOptional: `additionalContext` for hints." }, "typeVersion": 1 }, { "id": "e5d9ff12-601d-451d-ba62-f5b5a6d868c6", "name": "Sticky \u2014 Text Extraction", "type": "n8n-nodes-base.stickyNote", "position": [ 752, -32 ], "parameters": { "color": 7, "width": 356, "height": 502, "content": "## 2 \u00b7 Text Extraction\nRoute by file extension (.docx or .pdf) and extract plain text. Both paths feed into the AI identification step." }, "typeVersion": 1 }, { "id": "f55dcec1-3fb6-49e6-83b3-ff82d4f67c4c", "name": "Sticky \u2014 Identify", "type": "n8n-nodes-base.stickyNote", "position": [ 1120, -32 ], "parameters": { "color": 7, "width": 312, "height": 496, "content": "## 3 \u00b7 Identify Document Type\nGPT-4o reads the document and returns a structured JSON object containing the document type name and the full list of variable field placeholders specific to that document type." }, "typeVersion": 1 }, { "id": "1fd8e8e2-92de-4f55-b99c-3f926cfaf393", "name": "Sticky \u2014 Parse Classification", "type": "n8n-nodes-base.stickyNote", "position": [ 1456, -32 ], "parameters": { "color": 7, "width": 208, "height": 502, "content": "## 4 \u00b7 Parse AI Classification\nExtracts `documentType` and `fields` array from the JSON returned by the Identify step. Passes both forward alongside the original document text." }, "typeVersion": 1 }, { "id": "50a30ab0-229e-4480-a065-e23b7eddd6f4", "name": "Sticky \u2014 Generate Template", "type": "n8n-nodes-base.stickyNote", "position": [ 1680, -32 ], "parameters": { "color": 7, "width": 280, "height": 496, "content": "## 5 \u00b7 Generate Template\nA second GPT-4o pass uses the identified document type and field list to replace all variable content with [BRACKET] placeholders. Boilerplate and structure are preserved verbatim." }, "typeVersion": 1 }, { "id": "d454844d-9ce3-42bb-aa04-0643157842ee", "name": "Sticky \u2014 Build and Deliver", "type": "n8n-nodes-base.stickyNote", "position": [ 1968, -32 ], "parameters": { "color": 7, "width": 916, "height": 486, "content": "## 6 \u00b7 Build, Convert & Deliver\nClean AI output, wrap in HTML with standard letter formatting, render to PDF via Gotenberg, upload to Google Drive, set public viewer link, and return JSON response with file URLs." }, "typeVersion": 1 }, { "id": "73d3b058-e22f-4b90-beae-db324d40987e", "name": "Webhook \u2014 Receive Document", "type": "n8n-nodes-base.webhook", "position": [ 560, 176 ], "parameters": { "path": "general-template-creator", "options": {}, "httpMethod": "POST", "responseMode": "responseNode" }, "typeVersion": 2.1 }, { "id": "73f17329-0307-43e6-b09f-a995fe236796", "name": "Switch \u2014 Route by File Type", "type": "n8n-nodes-base.switch", "position": [ 800, 176 ], "parameters": { "rules": { "values": [ { "conditions": { "options": { "version": 2, "leftValue": "", "caseSensitive": true, "typeValidation": "strict" }, "combinator": "and", "conditions": [ { "id": "rule-docx", "operator": { "type": "string", "operation": "endsWith" }, "leftValue": "={{ $json.body.fileName.toLowerCase() }}", "rightValue": ".docx" } ] } }, { "conditions": { "options": { "version": 2, "leftValue": "", "caseSensitive": true, "typeValidation": "strict" }, "combinator": "and", "conditions": [ { "id": "rule-pdf", "operator": { "type": "string", "operation": "endsWith" }, "leftValue": "={{ $json.body.fileName.toLowerCase() }}", "rightValue": ".pdf" } ] } } ] }, "options": {} }, "typeVersion": 3.2 }, { "id": "bd11b29d-13cb-49ed-a0d4-7aba17260c8d", "name": "Extract \u2014 Word to Text", "type": "n8n-nodes-word2text.word2text", "position": [ 960, 96 ], "parameters": { "binaryPropertyName": "file" }, "typeVersion": 1 }, { "id": "a8c97b56-5cc3-429a-bb13-3be0075d3e47", "name": "Extract \u2014 PDF to Text", "type": "n8n-nodes-base.extractFromFile", "position": [ 960, 256 ], "parameters": { "options": {}, "operation": "pdf", "binaryPropertyName": "file" }, "typeVersion": 1.1 }, { "id": "6936cf79-77f9-4c48-934a-c51bbaef5ad7", "name": "AI \u2014 Identify Document Type", "type": "@n8n/n8n-nodes-langchain.agent", "position": [ 1152, 176 ], "parameters": { "text": "={{ $json.text }}", "options": { "systemMessage": "You are a document classification expert. Your sole job is to read a document and output a JSON object \u2014 nothing else.\n\nAnalyze the document and return ONLY a valid JSON object in this exact format:\n{\n \"documentType\": \"\",\n \"documentTypeSlug\": \"\",\n \"description\": \"\",\n \"fields\": [\n {\n \"placeholder\": \"[FIELD_NAME]\",\n \"description\": \"Brief description of what value goes here\"\n }\n ]\n}\n\nRules:\n- The fields array must contain EVERY variable element found in this specific document\n- Use ALL_CAPS_UNDERSCORE format for placeholder names inside brackets\n- Be specific: prefer [EMPLOYEE_NAME] over [NAME], [MONTHLY_RENT] over [AMOUNT]\n- Include ALL parties, ALL dates, ALL amounts, ALL addresses, ALL identifying numbers\n- Output ONLY the JSON object. No preamble, no explanation, no markdown fences." }, "promptType": "define" }, "typeVersion": 3.1 }, { "id": "291052c3-11e7-4e7c-a42c-668073e1be15", "name": "LLM \u2014 GPT-4o (Identify)", "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi", "position": [ 1152, 320 ], "parameters": { "model": { "__rl": true, "mode": "list", "value": "gpt-4o", "cachedResultName": "gpt-4o" }, "options": { "maxTokens": 1500, "temperature": 0.1 }, "builtInTools": {} }, "credentials": { "openAiApi": { "name": "" } }, "typeVersion": 1.3 }, { "id": "66ef38c5-dca4-4188-a208-20a3cf571025", "name": "Code \u2014 Parse Classification", "type": "n8n-nodes-base.code", "position": [ 1504, 176 ], "parameters": { "jsCode": "// Pull the AI output and original webhook data\nconst rawOutput = $('AI \u2014 Identify Document Type').item.json.output || '';\nconst webhookBody = $('Webhook \u2014 Receive Document').item.json.body || {};\nconst extractedText = $('AI \u2014 Identify Document Type').item.json.output ? $input.item.json.output : '';\n\n// Strip markdown fences if the model wrapped it anyway\nlet cleaned = rawOutput.replace(/```json\\s*/gi, '').replace(/```\\s*/g, '').trim();\n\nlet classification;\ntry {\n classification = JSON.parse(cleaned);\n} catch (e) {\n // Attempt to extract JSON object from within the string\n const match = cleaned.match(/\\{[\\s\\S]*\\}/);\n if (match) {\n classification = JSON.parse(match[0]);\n } else {\n throw new Error('Could not parse classification JSON from AI output: ' + cleaned.substring(0, 200));\n }\n}\n\n// Build a readable placeholder list for the template prompt\nconst fieldList = (classification.fields || []).map(f => f.placeholder + ' \u2014 ' + f.description).join('\\n');\n\nreturn [{\n json: {\n documentType: classification.documentType || 'Document',\n documentTypeSlug: classification.documentTypeSlug || 'document',\n documentDescription: classification.description || '',\n fields: classification.fields || [],\n fieldList: fieldList,\n fileName: webhookBody.fileName || 'unknown',\n additionalContext: webhookBody.additionalContext || ''\n }\n}];" }, "typeVersion": 2 }, { "id": "caa6e7da-e76c-4561-8820-cbf9a7fd09ce", "name": "AI \u2014 Generate Template", "type": "@n8n/n8n-nodes-langchain.agent", "position": [ 1712, 176 ], "parameters": { "text": "=You are an expert document template creator.\n\nYour task is to transform the provided document into a professional, reusable fill-in-the-blank template.\n\n**DOCUMENT TYPE:** {{ $json.documentType }}\n**DOCUMENT DESCRIPTION:** {{ $json.documentDescription }}\n**ORIGINAL FILENAME:** {{ $json.fileName }}\n\n**IDENTIFIED VARIABLE FIELDS FOR THIS DOCUMENT TYPE:**\n{{ $json.fieldList }}\n\n**DOCUMENT TEXT TO TEMPLATIZE:**\n{{ $('Extract \u2014 Word to Text').item.json.text || $('Extract \u2014 PDF to Text').item.json.text }}\n\n**YOUR INSTRUCTIONS:**\n1. Read the document carefully\n2. Replace ALL variable, instance-specific content with the appropriate [PLACEHOLDER] from the field list above\n3. If you encounter variable content NOT covered by the field list, create a new descriptive placeholder in [ALL_CAPS_UNDERSCORE] format\n4. Preserve ALL static boilerplate, standard language, and structural elements EXACTLY as written\n5. Maintain all section headings, numbering, indentation, and formatting conventions\n6. Keep signature blocks \u2014 replace signer names with [SIGNER_NAME] / [SIGNER_TITLE] placeholders\n\n**CRITICAL OUTPUT REQUIREMENTS:**\n- Output ONLY the template document text\n- Do NOT include any preamble, explanation, field list, or notes\n- Do NOT use markdown formatting (no **, no ```, no ###)\n- Do NOT add headers like \"TEMPLATE:\" or \"OUTPUT:\"\n- Start directly with the first line of the document\n- Use plain text only \u2014 output will be converted to PDF\n\nBegin output now.", "options": { "systemMessage": "You are a document template specialist. You convert real documents into reusable templates by replacing all variable, instance-specific content with clearly labeled [BRACKET] placeholders. You never alter boilerplate language, standard clauses, or the structural integrity of the document. You always output only the template document \u2014 no explanations, no metadata." }, "promptType": "define" }, "typeVersion": 3.1 }, { "id": "dd580e69-a34a-4161-ae8e-84ffeb2120a7", "name": "LLM \u2014 GPT-4o (Generate)", "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi", "position": [ 1712, 320 ], "parameters": { "model": { "__rl": true, "mode": "list", "value": "gpt-4o", "cachedResultName": "gpt-4o" }, "options": { "maxTokens": 4000, "temperature": 0.2 }, "builtInTools": {} }, "credentials": { "openAiApi": { "name": "" } }, "typeVersion": 1.3 }, { "id": "9f99c51b-1291-4ee5-81d7-dc653d87af03", "name": "Code \u2014 Clean and Format Output", "type": "n8n-nodes-base.code", "position": [ 2016, 176 ], "parameters": { "jsCode": "const rawAiOutput = $('AI \u2014 Generate Template').item.json.output || '';\nconst meta = $('Code \u2014 Parse Classification').item.json;\n\nlet cleanOutput = rawAiOutput;\n\n// Strip markdown code fences\ncleanOutput = cleanOutput.replace(/```[a-z]*\\n?/gi, '');\ncleanOutput = cleanOutput.replace(/```\\s*$/g, '');\n\n// Strip any preamble headers the model might add\ncleanOutput = cleanOutput.replace(/^(TEMPLATE:|OUTPUT:|DOCUMENT TEMPLATE:)\\s*\\n/gim, '');\ncleanOutput = cleanOutput.replace(/^\\*\\*(TEMPLATE|OUTPUT|RESULT):\\*\\*\\s*\\n/gim, '');\n\n// Strip placeholder/field lists if the model included them before the doc\ncleanOutput = cleanOutput.replace(/^[\\s\\S]*?(?=^[A-Z\\[\\(])/m, '');\n\n// Remove markdown bold\ncleanOutput = cleanOutput.replace(/\\*\\*([^*]+)\\*\\*/g, '$1');\n\n// Normalize whitespace\ncleanOutput = cleanOutput.replace(/\\n{3,}/g, '\\n\\n').trim();\n\n// Build filename\nconst timestamp = new Date().toISOString().split('T')[0];\nconst filename = meta.documentTypeSlug + '_template_' + timestamp;\n\n// Build HTML for Gotenberg\nconst htmlBody = cleanOutput.replace(/\\n/g, '
');\n\nconst htmlContent = '\\n\\n\\n \\n \\n\\n\\n ' + htmlBody + '\\n\\n';\n\nconst binaryData = Buffer.from(htmlContent, 'utf-8');\n\nreturn [{\n json: {\n filename: filename,\n documentType: meta.documentType,\n documentTypeSlug: meta.documentTypeSlug,\n documentDescription: meta.documentDescription,\n fields: meta.fields,\n originalFileName: meta.fileName,\n createdDate: timestamp,\n htmlContent: htmlContent,\n originalLength: rawAiOutput.length,\n cleanedLength: cleanOutput.length\n },\n binary: {\n data: {\n data: binaryData.toString('base64'),\n mimeType: 'text/html',\n fileName: 'index.html'\n }\n }\n}];" }, "typeVersion": 2 }, { "id": "15c8d810-ce67-4d11-9abe-bc3dcd00f031", "name": "Gotenberg \u2014 HTML to PDF", "type": "n8n-nodes-base.httpRequest", "position": [ 2192, 176 ], "parameters": { "method": "POST", "options": { "response": { "response": { "responseFormat": "file" } } }, "sendBody": true, "contentType": "multipart-form-data", "sendHeaders": true, "authentication": "genericCredentialType", "bodyParameters": { "parameters": [ { "name": "index.html", "parameterType": "formBinaryData", "inputDataFieldName": "=data" } ] }, "genericAuthType": "httpBasicAuth", "headerParameters": { "parameters": [ { "name": "Gotenberg-Output-Filename", "value": "={{ $('Code \u2014 Clean and Format Output').item.json.filename }}" } ] } }, "credentials": { "httpBasicAuth": { "name": "" } }, "typeVersion": 4.4 }, { "id": "a3fb5e14-a5d0-44a7-9c63-15e95fc57e09", "name": "GDrive \u2014 Upload PDF", "type": "n8n-nodes-base.googleDrive", "position": [ 2384, 176 ], "parameters": { "name": "={{ $('Code \u2014 Clean and Format Output').item.json.filename }}", "driveId": { "__rl": true, "mode": "list", "value": "My Drive" }, "options": {}, "folderId": { "__rl": true, "mode": "list", "value": "16p8nurkJi1eog833Z5lsaKc6zMtvn6rL", "cachedResultUrl": "https://drive.google.com/drive/folders/16p8nurkJi1eog833Z5lsaKc6zMtvn6rL", "cachedResultName": "Templates" } }, "credentials": { "googleDriveOAuth2Api": { "name": "" } }, "typeVersion": 3 }, { "id": "ae2e1e61-5e83-4b2e-9b45-9951421089b4", "name": "GDrive \u2014 Set Public Link", "type": "n8n-nodes-base.googleDrive", "position": [ 2576, 176 ], "parameters": { "fileId": { "__rl": true, "mode": "id", "value": "={{ $json.id }}" }, "options": {}, "operation": "share", "permissionsUi": { "permissionsValues": { "role": "reader", "type": "anyone" } } }, "credentials": { "googleDriveOAuth2Api": { "name": "" } }, "typeVersion": 3 }, { "id": "6baefca1-8050-48d9-b384-282ded4327ef", "name": "Webhook \u2014 Send Response", "type": "n8n-nodes-base.respondToWebhook", "position": [ 2752, 176 ], "parameters": { "options": { "responseHeaders": { "entries": [ { "name": "Content-Type", "value": "application/json" }, { "name": "Access-Control-Allow-Origin", "value": "*" } ] } }, "respondWith": "json", "responseBody": "={\n \"success\": true,\n \"message\": \"Document template created successfully!\",\n \"classification\": {\n \"documentType\": \"{{ $('Code \u2014 Clean and Format Output').item.json.documentType }}\",\n \"description\": \"{{ $('Code \u2014 Clean and Format Output').item.json.documentDescription }}\",\n \"fieldsIdentified\": {{ $('Code \u2014 Clean and Format Output').item.json.fields.length }}\n },\n \"template\": {\n \"filename\": \"{{ $('Code \u2014 Clean and Format Output').item.json.filename }}.pdf\",\n \"originalDocument\": \"{{ $('Code \u2014 Clean and Format Output').item.json.originalFileName }}\",\n \"createdDate\": \"{{ $('Code \u2014 Clean and Format Output').item.json.createdDate }}\",\n \"googleDriveId\": \"{{ $('GDrive \u2014 Upload PDF').item.json.id }}\",\n \"viewUrl\": \"{{ $('GDrive \u2014 Upload PDF').item.json.webViewLink }}\",\n \"downloadUrl\": \"{{ $('GDrive \u2014 Upload PDF').item.json.webContentLink }}\",\n \"directDownload\": \"https://drive.google.com/uc?export=download&id={{ $('GDrive \u2014 Upload PDF').item.json.id }}\"\n },\n \"fields\": {{ JSON.stringify($('Code \u2014 Clean and Format Output').item.json.fields) }},\n \"instructions\": {\n \"usage\": \"Your document template is ready. All variable content has been replaced with [PLACEHOLDER] tags.\",\n \"howToUse\": [\n \"Download the PDF to see the complete template with all placeholders\",\n \"Use the fields array in this response as your fill-in guide\",\n \"Replace each [PLACEHOLDER] with the actual value for your document\",\n \"Save as a new file for each new instance of the document\"\n ],\n \"placeholderFormat\": \"All placeholders use [ALL_CAPS_UNDERSCORE] format\"\n }\n}" }, "typeVersion": 1.5 } ], "active": false, "settings": { "binaryMode": "separate", "executionOrder": "v1" }, "versionId": "b872b98b-affd-41f5-8068-53e511494a91", "connections": { "GDrive \u2014 Upload PDF": { "main": [ [ { "node": "GDrive \u2014 Set Public Link", "type": "main", "index": 0 } ] ] }, "Extract \u2014 PDF to Text": { "main": [ [ { "node": "AI \u2014 Identify Document Type", "type": "main", "index": 0 } ] ] }, "AI \u2014 Generate Template": { "main": [ [ { "node": "Code \u2014 Clean and Format Output", "type": "main", "index": 0 } ] ] }, "Extract \u2014 Word to Text": { "main": [ [ { "node": "AI \u2014 Identify Document Type", "type": "main", "index": 0 } ] ] }, "Gotenberg \u2014 HTML to PDF": { "main": [ [ { "node": "GDrive \u2014 Upload PDF", "type": "main", "index": 0 } ] ] }, "LLM \u2014 GPT-4o (Generate)": { "ai_languageModel": [ [ { "node": "AI \u2014 Generate Template", "type": "ai_languageModel", "index": 0 } ] ] }, "LLM \u2014 GPT-4o (Identify)": { "ai_languageModel": [ [ { "node": "AI \u2014 Identify Document Type", "type": "ai_languageModel", "index": 0 } ] ] }, "GDrive \u2014 Set Public Link": { "main": [ [ { "node": "Webhook \u2014 Send Response", "type": "main", "index": 0 } ] ] }, "Webhook \u2014 Receive Document": { "main": [ [ { "node": "Switch \u2014 Route by File Type", "type": "main", "index": 0 } ] ] }, "AI \u2014 Identify Document Type": { "main": [ [ { "node": "Code \u2014 Parse Classification", "type": "main", "index": 0 } ] ] }, "Code \u2014 Parse Classification": { "main": [ [ { "node": "AI \u2014 Generate Template", "type": "main", "index": 0 } ] ] }, "Switch \u2014 Route by File Type": { "main": [ [ { "node": "Extract \u2014 Word to Text", "type": "main", "index": 0 } ], [ { "node": "Extract \u2014 PDF to Text", "type": "main", "index": 0 } ] ] }, "Code \u2014 Clean and Format Output": { "main": [ [ { "node": "Gotenberg \u2014 HTML to PDF", "type": "main", "index": 0 } ] ] } } }