{ "meta": { "templateCredsSetupCompleted": true }, "nodes": [ { "id": "7a22b325-5cc4-4410-a3e0-4b6157bf2fd1", "name": "Universal Safe Parser", "type": "n8n-nodes-base.code", "position": [ 784, 3632 ], "parameters": { "jsCode": "// === UNIVERSAL SAFE PARSER (STANDARDIZATION + UNIQUE ID) ===\nconst results = [];\n\nfor (const item of $input.all()) {\n const json = item.json || {};\n const html = json.html || json.data || json.content || \"\";\n const sourceUrl =\n $(\"Loop Through Sources (Rate Limit Safe)\").first().json.source_url || \"\";\n let sourceType = $(\"Loop Through Sources (Rate Limit Safe)\").first().json\n .source_type;\n let productName = $(\"Loop Through Sources (Rate Limit Safe)\").first().json\n .product_name;\n\n if (!sourceType || sourceType === \"unknown\") {\n sourceType = detectSource(sourceUrl, html);\n }\n\n /** @type {any[]} */\n let reviews = [];\n let error = null;\n\n if (!html || html.length < 200) {\n error = \"No HTML content\";\n } else {\n try {\n if (sourceType === \"trustpilot\") reviews = parseTrustpilot(html);\n else if (sourceType === \"capterra\") reviews = parseCapterra(html);\n else if (sourceType === \"chromewebstore\")\n reviews = parseChromeWebStore(html);\n else if (sourceType === \"producthunt\") reviews = parseProductHunt(html);\n else if (sourceType === \"trustradius\") reviews = parseTrustRadius(html);\n else error = `Unknown source: ${sourceType}`;\n } catch (e) {\n error = e instanceof Error ? e.message : String(e);\n //error = e.message;\n }\n }\n\n // === STANDARDIZE EVERY REVIEW + ADD UNIQUE ID ===\n const standardizedReviews = (reviews || []).map((review) => {\n const cleanText = (review.text || \"\").trim();\n const cleanTitle = (review.title || \"\").trim();\n const author = (review.author || \"Anonymous\").trim();\n let date = null;\n if (review.date) {\n const d = new Date(review.date);\n // Check if 'd' is a valid date object\n if (!isNaN(d.getTime())) {\n try {\n date = d.toISOString().split(\"T\")[0];\n } catch (e) {\n date = null; // Fallback if ISO conversion fails\n }\n } else {\n // Optional: If date is \"2 days ago\", it ends up here.\n // You can leave it null or map it to today's date if strictly necessary.\n date = null;\n }\n }\n\n // === UNIQUE ID: deterministic hash based on source + author + date + first 100 chars of text ===\n // This survives reformatting, extra spaces, etc. \u2014 99.99% duplicate-proof\n const hashString = `${sourceType}|${author}|${date}|${cleanText.substring(0, 150)}`;\n function simpleHash(str) {\n let hash = 0;\n for (let i = 0; i < str.length; i++) {\n const char = str.charCodeAt(i);\n hash = (hash << 5) - hash + char;\n hash = hash & hash; // Convert to 32-bit integer\n }\n return Math.abs(hash).toString(36); // short, URL-safe string\n }\n const unique_id = `${sourceType}_${simpleHash(hashString)}`;\n\n return {\n unique_id, // \u2190 Your primary key / dedupe field\n author: author,\n rating: review.rating ? Number(review.rating) : null,\n title: cleanTitle || \"[No title provided]\",\n text: cleanText,\n pros: review.pros || \"[Not specified]\",\n cons: review.cons || \"[Not specified]\",\n date: date,\n job_title: review.job_title || \"[Not provided]\",\n company_size: review.company_size || \"[Unknown]\",\n industry: review.industry || \"[Unknown]\",\n location: review.location || \"[Not specified]\",\n verified: review.verified || false,\n\n scraped_at: new Date().toISOString(),\n };\n });\n\n const previousPage = $(\"URL Builder\").first().json.page_number || 1;\n results.push({\n json: {\n //...json,\n // === CRITICAL: PASS METADATA FOR PAGINATION LOOP ===\n product_name: productName,\n page_number: previousPage,\n review_count: standardizedReviews.length,\n reviews: standardizedReviews,\n source_url: sourceUrl,\n source_type: sourceType, \n success: !error,\n error_message: error,\n },\n });\n}\n\n // === STEP 2: SAVE TO GLOBAL STORAGE ===\n const staticData = $getWorkflowStaticData('global');\n // Initialize if it doesn't exist (safety check)\n if (!staticData.allScrapedReviews) staticData.allScrapedReviews = [];\n \n // Append the results from this specific loop iteration to the global list\n // We map 'results' because it contains { json: ... } structure\n const cleanItems = results.map(r => r.json);\n staticData.allScrapedReviews = staticData.allScrapedReviews.concat(cleanItems);\n\nreturn results;\n\n// ==================== HELPER FUNCTIONS ====================\n\nfunction detectSource(url, html) {\n const u = (url || \"\").toLowerCase();\n const h = html || \"\";\n\n // A. Check URL first\n if (u.includes(\"trustpilot\")) return \"trustpilot\";\n if (u.includes(\"capterra\")) return \"capterra\";\n if (u.includes(\"chrome.google.com\") || u.includes(\"chromewebstore\"))\n return \"chromewebstore\";\n if (u.includes(\"producthunt\")) return \"producthunt\";\n if (u.includes(\"trustradius\")) return \"trustradius\";\n\n // B. Check HTML Signatures (Fallback)\n if (\n h.includes(\"data-service-review-card-paper\") ||\n h.includes(\"trustpilot-widget\")\n )\n return \"trustpilot\";\n if (h.includes(\"__next_f\") && h.includes(\"CapterraProduct\"))\n return \"capterra\";\n if (h.includes(\"T7rvce\") && h.includes(\"LfYwpe\")) return \"chromewebstore\";\n if (\n h.includes(\"DetailedReviewConnection\") ||\n h.includes(\"DetailedReviewEdge\")\n )\n return \"producthunt\";\n if (h.includes(\"Review_review__\") || h.includes(\"trScore\"))\n return \"trustradius\";\n\n return \"unknown\";\n}\n\nfunction decodeHtml(html) {\n if (!html) return \"\";\n return html\n .replace(/&/g, \"&\")\n .replace(/</g, \"<\")\n .replace(/>/g, \">\")\n .replace(/"/g, '\"')\n .replace(/'/g, \"'\")\n .replace(//gi, \"\\n\");\n}\n\n// --- 1. Trustpilot Parser ---\nfunction parseTrustpilot(html) {\n if (\n html.includes(\"cf-browser-verification\") ||\n html.includes(\"Just a moment\")\n ) {\n throw new Error(\"Blocked by Cloudflare.\");\n }\n const cards = html.split('data-service-review-card-paper=\"true\"').slice(1);\n const extracted = [];\n\n for (const card of cards) {\n const authorMatch = card.match(\n /data-consumer-name-typography=\"true\">([^<]+)([^<]+)([\\s\\S]*?)(?=<\\/p>|([^<]+)self\\.__next_f\\.push\\(\\[1,\"([\\s\\S]*?)\"\\]\\)<\\/script>/g;\n let match;\n while ((match = regex.exec(html)) !== null) {\n fullDataString += match[1];\n }\n\n if (!fullDataString) {\n throw new Error(\"No Next.js data script tags found.\");\n }\n\n // C. Build Data Map (Your Split Logic)\n const dataMap = new Map();\n // In the regex capture, newlines are stored as the literal string \"\\\\n\"\n const chunks = fullDataString.split(\"\\\\n\");\n\n chunks.forEach((chunk) => {\n if (!chunk) return;\n const firstColonIndex = chunk.indexOf(\":\");\n if (firstColonIndex === -1) return;\n\n const key = chunk.substring(0, firstColonIndex);\n let value = chunk.substring(firstColonIndex + 1);\n // Unescape quotes *after* splitting\n value = value.replace(/\\\\\"/g, '\"');\n dataMap.set(key, value);\n });\n\n if (dataMap.size === 0)\n throw new Error(\"Data map empty. Regex succeeded but split failed.\");\n\n // D. Recursive Resolver (Your Cache Logic)\n const resolvedCache = new Map();\n function resolve(key) {\n if (resolvedCache.has(key)) return resolvedCache.get(key);\n\n let valueStr = dataMap.get(key);\n if (!valueStr) return null;\n\n let result;\n if (valueStr.startsWith(\"{\") || valueStr.startsWith(\"[\")) {\n try {\n result = JSON.parse(valueStr, (reviverKey, reviverValue) => {\n if (\n typeof reviverValue === \"string\" &&\n reviverValue.startsWith(\"$\")\n ) {\n const refKey = reviverValue.substring(1);\n return resolve(refKey);\n }\n return reviverValue;\n });\n } catch (e) {\n result = null;\n }\n } else if (valueStr.startsWith(\"$\")) {\n result = resolve(valueStr.substring(1));\n } else if (valueStr.startsWith('\"') && valueStr.endsWith('\"')) {\n result = valueStr.substring(1, valueStr.length - 1);\n } else {\n result = valueStr;\n }\n resolvedCache.set(key, result);\n return result;\n }\n\n // E. Find Main Reviews Key (Your Key Finder)\n let mainReviewsKey = null;\n for (const [key, value] of dataMap.entries()) {\n if (\n value.includes('\"__typename\":\"CapterraProduct\"') &&\n value.includes('\"textReviews\":')\n ) {\n const reviewKeyMatch = value.match(\n /\"textReviews\":\"\\$([a-zA-Z0-9]{1,5})\"/,\n );\n if (reviewKeyMatch && reviewKeyMatch[1]) {\n mainReviewsKey = reviewKeyMatch[1];\n break;\n }\n }\n }\n\n // Fallback strategy if Product Object search fails\n if (!mainReviewsKey) {\n for (const [key, value] of dataMap.entries()) {\n if (value.startsWith('[\"$') && value.endsWith('\"]')) {\n // Basic heuristic: assume the first array of references might be it\n mainReviewsKey = key;\n break;\n }\n }\n }\n\n if (!mainReviewsKey) throw new Error(\"Could not find main review list key.\");\n\n // F. Resolve List\n const mainReviewsList = resolve(mainReviewsKey);\n if (!Array.isArray(mainReviewsList))\n throw new Error(\"Resolved review list is not an array.\");\n\n // G. Map Data (Your Mapper)\n const getRating = (rating) =>\n rating !== null && rating !== undefined\n ? parseFloat(rating).toString()\n : null;\n\n return mainReviewsList\n .map((review) => {\n if (!review || !review.reviewer) return null;\n\n let formattedDate = review.writtenOn; // Keep original if parsing fails\n try {\n // Simple date formatting\n formattedDate = new Date(review.writtenOn).toISOString().split(\"T\")[0];\n } catch (e) {}\n\n return {\n author: review.reviewer.anonymityOn\n ? \"Verified Reviewer\"\n : review.reviewer.fullName,\n job_title: review.reviewer.jobTitle,\n company_size: review.reviewer.companySize,\n industry: review.reviewer.industry,\n rating: getRating(review.overallRating),\n title: review.title,\n text: review.generalComments,\n pros: review.prosText,\n cons: review.consText,\n date: formattedDate,\n verified: review.reviewer.isValidated ? \"Yes\" : \"No\",\n source: \"capterra\",\n };\n })\n .filter((r) => r !== null);\n}\n\n// --- 3. Chrome Web Store Parser ---\nfunction parseChromeWebStore(html) {\n const chunks = html.split('class=\"T7rvce\"').slice(1);\n const extracted = [];\n for (const chunk of chunks) {\n const authorMatch = chunk.match(/class=\"LfYwpe\">([^<]+)([^<]+)]*>([\\s\\S]*?)<\\/p>/);\n\n if (authorMatch) {\n extracted.push({\n author: authorMatch[1].trim(),\n rating: ratingMatch ? parseInt(ratingMatch[1]) : null,\n date: dateMatch ? dateMatch[1].trim() : null,\n text: textMatch ? decodeHtml(textMatch[1]).trim() : null,\n source: \"chromewebstore\",\n });\n }\n }\n return extracted;\n}\n\n// --- 4. Product Hunt Parser ---\nfunction parseProductHunt(html) {\n const extracted = [];\n const seenIds = new Set();\n \n const startMatch = html.match(/DetailedReviewConnection.*?edges/);\n if (!startMatch) return []; // Return empty if not found\n \n const section = html.substring(startMatch.index);\n const blocks = section.split(/DetailedReviewEdge.*?node.*?DetailedReview/);\n blocks.shift();\n\n const getVal = (block, key) => {\n const r = new RegExp(`\"${key}\":\"([^\"]+)\"`);\n const m = block.match(r);\n return m ? m[1] : null;\n };\n\n for (const block of blocks) {\n const reviewId = getVal(block, \"id\");\n if (reviewId && seenIds.has(reviewId)) {\n continue;\n }\n if (reviewId) seenIds.add(reviewId);\n \n const name = getVal(block, \"name\") || \"Anonymous\";\n const body = getVal(block, \"body\") || getVal(block, \"overallExperience\");\n const date = getVal(block, \"createdAt\");\n\n if (body) {\n extracted.push({\n unique_id: `producthunt_${reviewId}`,\n author: name,\n text: decodeHtml(body.replace(/\\\\n/g, \"\\n\")),\n date: date,\n source: \"producthunt\",\n });\n }\n }\n return extracted;\n}\n\n// --- 5. TrustRadius Parser ---\nfunction parseTrustRadius(html) {\n const blocks = html.split('
]+>]+>([^<]+)<\\/a>/,\n );\n const ratingMatch = block.match(/data-rating=\"([\\d\\.]+)\"/);\n const nameMatch = block.match(/Byline_name__[^>]+>([^<]+)<\\/div>/);\n const dateMatch = block.match(/datetime=\"([^\"]+)\"/);\n const summaryMatch = block.match(\n /Use Cases and Deployment Scope<\\/h3>]*>(.*?)<\\/p>/,\n );\n\n if (ratingMatch) {\n extracted.push({\n author: nameMatch ? decodeHtml(nameMatch[1]) : \"Anonymous\",\n rating: parseFloat(ratingMatch[1]),\n title: titleMatch ? decodeHtml(titleMatch[1]) : null,\n text: summaryMatch ? decodeHtml(summaryMatch[1]) : null,\n date: dateMatch ? dateMatch[1] : null,\n source: \"trustradius\",\n });\n }\n }\n return extracted;\n}\n" }, "typeVersion": 2, "alwaysOutputData": true }, { "id": "a3ebf41d-973d-4426-800a-82454b59cd56", "name": "URL Builder", "type": "n8n-nodes-base.code", "position": [ 320, 3632 ], "parameters": { "jsCode": "// === DYNAMIC URL BUILDER ===\nconst item = $input.first().json;\n\n// Get the current page number from the loop context, default to 1\nconst page = item.page_number || 1;\nlet url = item.source_url;\nconst type = item.source_type;\n\nif (!url) {\n throw new Error(\"Pagination Error: 'source_url' was lost in the loop.\");\n}\n\n// 1. Chrome Web Store: Skip pagination (Infinite Scroll logic is too complex for this method)\nif (type === 'chromewebstore') {\n return { ...item, target_url: url };\n}\n\n// 2. Pagination Logic\n// If we are on Page 1, we usually don't need the parameter (safer to avoid redirects)\n\nif (page > 1) {\n // If ProductHunt already has params (like ?feed=single), this correctly uses '&'\n const separator = url.includes('?') ? '&' : '?';\n url = `${url}${separator}page=${page}`;\n}\n\nreturn { \n ...item, \n target_url: url \n};" }, "typeVersion": 2 }, { "id": "6a3780d0-3d41-4572-8fbd-09aa88e15669", "name": "Manual Trigger", "type": "n8n-nodes-base.manualTrigger", "position": [ -688, 3616 ], "parameters": {}, "typeVersion": 1 }, { "id": "33b6d69b-4e91-4d18-af3f-71f1ddfaa4ae", "name": "Prepare Review Sources", "type": "n8n-nodes-base.code", "position": [ -352, 3616 ], "parameters": { "jsCode": "// === 1. Reset global storage for scraped reviews ===\nconst staticData = $getWorkflowStaticData('global');\nstaticData.allScrapedReviews = [];\n\n// === 2. Hard-coded defaults (your fallback) ===\nconst defaultProduct = {\n name: \"Thordata\",\n trustpilot_domain: \"thordata.com\",\n capterra_url: \"\",\n chrome_webstore_url: \"\",\n trustradius_url: \"\",\n producthunt_slug: \"\"\n};\n\n// === 3. Start with defaults ===\nlet product = { ...defaultProduct };\n\n// === 4. Override with Form input (highest priority) ===\nconst formData = $input.first()?.json;\n\nif (formData && (\n formData.productName || \n formData.domain || \n formData.capterraUrl || \n formData.chromeStoreUrl || \n formData.trustRadius || \n formData.producthunt\n)) {\n product.name = (formData.productName || defaultProduct.name)?.trim() || defaultProduct.name;\n product.trustpilot_domain = (formData.domain || defaultProduct.trustpilot_domain)?.trim() || defaultProduct.trustpilot_domain;\n product.capterra_url = (formData.capterraUrl || defaultProduct.capterra_url)?.trim() || \"\";\n product.chrome_webstore_url = (formData.chromeStoreUrl || defaultProduct.chrome_webstore_url)?.trim() || \"\";\n product.trustradius_url = (formData.trustRadius || defaultProduct.trustradius_url)?.trim() || \"\";\n product.producthunt_slug = (formData.producthunt || defaultProduct.producthunt_slug)?.trim() || \"\";\n}\n\n// === 5. Override with Webhook POST body (if present) ===\nelse if ($node[\"Webhook Trigger\"] && $input.first()?.json && Object.keys($input.first().json).length > 0) {\n const body = $input.first().json;\n \n product.name = (body.productName || defaultProduct.name)?.trim() || defaultProduct.name;\n product.trustpilot_domain = (body.domain || defaultProduct.trustpilot_domain)?.trim() || defaultProduct.trustpilot_domain;\n product.capterra_url = (body.capterraUrl || \"\").toString().trim() || \"\";\n product.chrome_webstore_url = (body.chromeStoreUrl || \"\").toString().trim() || \"\";\n product.trustradius_url = (body.trustRadius || \"\").toString().trim() || \"\";\n product.producthunt_slug = (body.producthunt || defaultProduct.producthunt_slug)?.trim() || \"\";\n}\n\n// (Manual \u201cExecute Workflow\u201d button \u2192 uses defaults only)\n\n// === 6. Build the standardized sources array ===\nconst sources = [];\n\nif (product.trustpilot_domain) {\n sources.push({\n product_name: product.name,\n source_type: \"trustpilot\",\n source_url: \"https://www.trustpilot.com/review/\" + product.trustpilot_domain\n });\n}\n\nif (product.capterra_url) {\n sources.push({\n product_name: product.name,\n source_type: \"capterra\",\n source_url: product.capterra_url\n });\n}\n\nif (product.chrome_webstore_url) {\n sources.push({\n product_name: product.name,\n source_type: \"chromewebstore\",\n source_url: product.chrome_webstore_url\n });\n}\n\nif (product.trustradius_url) {\n sources.push({\n product_name: product.name,\n source_type: \"trustradius\",\n source_url: product.trustradius_url\n });\n}\n\nif (product.producthunt_slug) {\n sources.push({\n product_name: product.name,\n source_type: \"producthunt\",\n source_url: \"https://www.producthunt.com/products/\" + product.producthunt_slug + \"/reviews?feed=single&filter=all&order=newest\"\n });\n}\n\n// === 7. Return one item per source ===\nreturn sources.map(source => ({ json: source }));" }, "typeVersion": 2 }, { "id": "202cbefc-0ac8-4f8f-b497-802f3598c0b1", "name": "Initialize Pagination", "type": "n8n-nodes-base.set", "position": [ 96, 3632 ], "parameters": { "options": {}, "assignments": { "assignments": [ { "id": "d12b0a88-1a31-4cff-9abf-dc25101b6de4", "name": "page_number", "type": "number", "value": 1 } ] }, "includeOtherFields": true }, "typeVersion": 3.4 }, { "id": "347afbbd-8f94-41aa-b0db-b159f70431a9", "name": "Loop Through Sources (Rate Limit Safe)", "type": "n8n-nodes-base.splitInBatches", "position": [ -128, 3616 ], "parameters": { "options": {} }, "typeVersion": 3 }, { "id": "5708ec7e-2a50-49b1-bbd5-dc2b514e29ae", "name": "Scrape Page (Robust) via Thordata API", "type": "n8n-nodes-base.httpRequest", "onError": "continueRegularOutput", "position": [ 560, 3632 ], "parameters": { "url": "https://universalapi.thordata.com/request", "method": "POST", "options": { "proxy": "http://td-customer-username:user@example.com:9999", "timeout": 120000 }, "sendBody": true, "contentType": "form-urlencoded", "authentication": "genericCredentialType", "bodyParameters": { "parameters": [ { "name": "url", "value": "={{ $json.target_url }}" }, { "name": "type", "value": "html" }, { "name": "js_render", "value": "True" }, { "name": "clean_content", "value": "css" } ] }, "genericAuthType": "httpHeaderAuth" }, "credentials": { "httpHeaderAuth": { "name": "" } }, "retryOnFail": true, "typeVersion": 4.3, "alwaysOutputData": true }, { "id": "fb1e5835-d23c-47fd-a4cc-54648f5b4817", "name": "Has More Pages?", "type": "n8n-nodes-base.if", "position": [ 1024, 3632 ], "parameters": { "options": {}, "conditions": { "options": { "version": 2, "leftValue": "", "caseSensitive": true, "typeValidation": "loose" }, "combinator": "or", "conditions": [ { "id": "d0cee53f-bdb6-4e2a-a9ac-e18950366991", "operator": { "type": "boolean", "operation": "equals" }, "leftValue": "={{ ($json.reviews.length > 0 || $json.review_count > 0) && !['chromewebstore', 'producthunt'].includes($json.source_type) }}", "rightValue": true } ] }, "looseTypeValidation": true }, "typeVersion": 2.2 }, { "id": "e2131ae6-aff2-47c8-8fbf-768f38696bc2", "name": "Increment Page Counter", "type": "n8n-nodes-base.set", "position": [ 1312, 3616 ], "parameters": { "options": {}, "assignments": { "assignments": [ { "id": "ac85baa8-149f-4c43-bb8e-58b44b39ecc0", "name": "page_number", "type": "number", "value": "={{ $json.page_number + 1 }}" } ] }, "includeOtherFields": true }, "typeVersion": 3.4 }, { "id": "e457c5c7-51df-491b-addb-03e718d3ddfb", "name": "Rate Limit Delay (2s)", "type": "n8n-nodes-base.wait", "position": [ 1520, 3616 ], "parameters": { "amount": 2 }, "typeVersion": 1.1 }, { "id": "c7af6943-9a5c-4d47-a1c6-4a18f76d8d40", "name": "Deduplicate & Summarize All Reviews", "type": "n8n-nodes-base.code", "position": [ 96, 3424 ], "parameters": { "jsCode": "// === STEP 3: READ GLOBAL STORAGE ===\nconst staticData = $getWorkflowStaticData('global');\nconst allItems = staticData.allScrapedReviews || [];\n\nconst allReviews = [];\nconst errors = [];\nconst seenIds = new Set(); // \u2190 Extra safety net\nconst dedupedReviews = [];\n\nallItems.forEach(item => {\n if (!item) return [];\n const json = item.json || item;\n if (!json) return [];\n \n const { reviews = [], success, error_message, source_detected, source_url } = json;\n\n if (!success) {\n errors.push({ source: source_detected || source_url, error: error_message });\n } else {\n // Ensure reviews is actually an array before looping\n if (Array.isArray(reviews)) { \n reviews.forEach(r => {\n if (!seenIds.has(r.unique_id)) {\n seenIds.add(r.unique_id);\n dedupedReviews.push(r);\n }\n });\n }\n }\n});\n\nreturn [{\n json: {\n summary: {\n product: \"Thordata\",\n total_reviews_scraped: dedupedReviews.length,\n unique_reviews: dedupedReviews.length,\n sources_attempted: $input.all().length,\n sources_failed: errors.length,\n scraped_at: new Date().toISOString()\n },\n reviews: dedupedReviews,\n errors: errors.length > 0 ? errors : null\n }\n}];" }, "typeVersion": 2 }, { "id": "021e2bfb-a706-4932-888a-c54f43db63a9", "name": "Prepare Data for AI Analysis", "type": "n8n-nodes-base.code", "position": [ 320, 3424 ], "parameters": { "jsCode": "const { summary, reviews } = $input.first().json;\n\n// Format reviews into readable text\nconst reviewText = reviews.map((r, idx) => {\n return `Review #${idx + 1} (${r.source} - Rating: ${r.rating}/5):\nAuthor: ${r.author || 'Anonymous'}\nDate: ${r.date}\n${r.title ? 'Title: ' + r.title : ''}\nText: ${r.text || '[No text provided]'}`;\n}).join('\\n\\n---\\n\\n');\n\n// Calculate rating distribution\nconst ratings = reviews.map(r => r.rating).filter(r => r);\nconst avgRating = ratings.length > 0 \n ? (ratings.reduce((a, b) => a + b, 0) / ratings.length).toFixed(2) \n : 0;\n\nconst ratingDist = {\n '5': ratings.filter(r => r === 5).length,\n '4': ratings.filter(r => r === 4).length,\n '3': ratings.filter(r => r === 3).length,\n '2': ratings.filter(r => r === 2).length,\n '1': ratings.filter(r => r === 1).length\n};\n\n// Sources breakdown\nconst sources = [...new Set(reviews.map(r => r.source))];\n\nreturn [{\n json: {\n // Flattened for template\n product: summary.product,\n total_reviews: summary.unique_reviews,\n avg_rating: avgRating,\n \n // For the prompt\n all_reviews_text: reviewText,\n \n // Additional context\n rating_distribution: ratingDist,\n sources: sources,\n date_range: {\n earliest: reviews.map(r => r.date).sort()[0],\n latest: reviews.map(r => r.date).sort().reverse()[0]\n },\n \n // Keep original for reference\n original_reviews: reviews,\n original_summary: summary\n }\n}];" }, "typeVersion": 2 }, { "id": "861af668-6cae-4e1e-9c72-1c29b959506e", "name": "AI Review Sentiment Analysis", "type": "@n8n/n8n-nodes-langchain.openAi", "position": [ 544, 3424 ], "parameters": { "modelId": { "__rl": true, "mode": "list", "value": "gpt-4.1-2025-04-14", "cachedResultName": "GPT-4.1-2025-04-14" }, "options": {}, "responses": { "values": [ { "content": "=You are a B2B product analyst. Analyze these {{$json.total_reviews}} product reviews collectively.\n\nPRODUCT: {{$json.product}}\nAVERAGE RATING: {{$json.avg_rating}}/5\nDATE RANGE: {{$json.date_range.earliest}} to {{$json.date_range.latest}}\nSOURCES: {{$json.sources}}\n\nRating Distribution:\n- 5 stars: {{$json.rating_distribution[\"5\"]}} reviews\n- 1 star: {{$json.rating_distribution[\"1\"]}} reviews\n\nREVIEWS:\n{{$json.all_reviews_text}}\n\nAnalyze all reviews collectively. Provide ONLY JSON, NO markdown:\n\n{\n \"collective_sentiment_score\": <-1 to 1>,\n \"collective_sentiment_label\": \"\",\n \"confidence\": <0-1>,\n \"overall_emotion\": \"\",\n \"urgency\": \"\",\n \n \"batch_analysis\": {\n \"common_praise\": [\"theme 1\", \"theme 2\", \"theme 3\"],\n \"common_complaints\": [\"issue 1\", \"issue 2\", \"issue 3\"],\n \"critical_issues\": [{\"issue\": \"description\", \"mentions\": 3, \"severity\": \"high\"}],\n \"key_features_mentioned\": [\"feature 1\", \"feature 2\"],\n \"pricing_sentiment\": \"\",\n \"support_quality\": \"\"\n },\n \n \"trend\": {\n \"direction\": \"\",\n \"key_finding\": \"main insight\"\n },\n \n \"recommended_actions\": {\n \"product_team\": \"action\",\n \"sales_team\": \"action\",\n \"support_team\": \"action\"\n },\n \n \"overall_health\": \"\",\n \"churn_risk\": \"\"\n}" } ] }, "builtInTools": {} }, "credentials": { "openAiApi": { "name": "" } }, "typeVersion": 2 }, { "id": "e0b090cb-ef8b-4b07-aa89-46df26731575", "name": "Return Final Results", "type": "n8n-nodes-base.respondToWebhook", "position": [ 656, 3152 ], "parameters": { "options": {}, "respondWith": "json", "responseBody": "={{\n {\n \"product_name\": $json.product,\n \"stats\": $json.original_summary,\n \"data\": $json.original_reviews\n }\n}}" }, "typeVersion": 1.4 }, { "id": "1bda3f48-85b5-4a18-b5e6-7bb4e1d866f0", "name": "Webhook Trigger", "type": "n8n-nodes-base.webhook", "position": [ -688, 3840 ], "parameters": { "path": "c35353c9-a0e5-4f0a-86a4-cc5466bd83fd", "options": {}, "httpMethod": "POST", "responseMode": "responseNode" }, "typeVersion": 2.1 }, { "id": "e0faf882-df2d-424f-956c-42bcf461b92c", "name": "Form Trigger \u2013 Submit Sources", "type": "n8n-nodes-base.formTrigger", "position": [ -688, 3408 ], "parameters": { "path": "sources", "options": {}, "formTitle": "Submit Product Review Sources", "formFields": { "values": [ { "fieldLabel": "productName", "placeholder": "Thordata", "requiredField": true }, { "fieldLabel": "domain", "placeholder": "thordata.com" }, { "fieldLabel": "capterraUrl", "placeholder": "https://www.capterra.com/p/168264/ZoomInfo-sales/reviews" }, { "fieldLabel": "chromeStoreUrl", "placeholder": "https://chromewebstore.google.com/detail/vpn-for-chrome-nordvpn-pr/YOUR_AWS_SECRET_KEY_HERE" }, { "fieldLabel": "trustRadius", "placeholder": "https://www.trustradius.com/products/leadgrabber-pro/reviews/all" }, { "fieldLabel": "producthunt", "placeholder": "gemini-6" } ] }, "formDescription": "Provide review source URLs for a product. Placeholders shown for convenience, replace with your product's details as needed." }, "typeVersion": 1 }, { "id": "1d1fd03e-0fdf-47a1-8773-b542a7ed192a", "name": "Render Elegant Email Template", "type": "n8n-nodes-base.code", "position": [ 912, 3424 ], "parameters": { "jsCode": "// === 1. Get the AI raw JSON string ===\nconst openAiOutput = $input.first().json.output?.[0]?.content?.[0]?.text;\n\nif (!openAiOutput) {\n throw new Error(\"OpenAI response missing\");\n}\n\nlet aiAnalysis;\ntry {\n aiAnalysis = JSON.parse(openAiOutput);\n} catch (e) {\n throw new Error(`Failed to parse AI response as JSON:\\n${openAiOutput}`);\n}\n\n// Now safe to use\nconst product = $('Prepare Data for AI Analysis').first().json.product || \"Unknown Product\";\nconst totalReviews = $('Prepare Data for AI Analysis').first().json.total_reviews || 0;\nconst avgRating = $('Prepare Data for AI Analysis').first().json.avg_rating ? Number($('Prepare Data for AI Analysis').first().json.avg_rating).toFixed(2) : \"N/A\";\n\n// === 3. Colors ===\nconst sentimentColor = {\n positive: \"#10b981\",\n neutral: \"#f59e0b\",\n negative: \"#ef4444\"\n}[aiAnalysis.collective_sentiment_label] || \"#6b7280\";\n\nconst healthColor = aiAnalysis.overall_health === \"healthy\" ? \"#10b981\" : \"#ef4444\";\nconst churnColor = aiAnalysis.churn_risk === \"low\" ? \"#10b981\" : \n aiAnalysis.churn_risk === \"medium\" ? \"#f59e0b\" : \"#ef4444\";\n\n// === 4. Generate stunning responsive HTML email ===\nconst html = `\n\n\n \n \n ${product} Review Intelligence Report\n \n \n\n\n
\n
\n

${product}

\n

AI-Powered Review Intelligence Report \u2022 ${new Date().toLocaleDateString('en-US', { weekday: 'long', year: 'numeric', month: 'long', day: 'numeric' })}

\n
\n\n
\n
\n
\n ${aiAnalysis.collective_sentiment_label.toUpperCase()} SENTIMENT\n
\n
\n\n
\n
\n
${totalReviews || 0}
\n
Reviews Analyzed
\n
\n
\n
${avgRating || 'N/A'}
\n
Average Rating
\n
\n
\n
${aiAnalysis.overall_health?.toUpperCase()}
\n
Product Health
\n
\n
\n
${aiAnalysis.churn_risk?.toUpperCase()}
\n
Churn Risk
\n
\n
\n\n
\n

Top Praises

\n
    ${(aiAnalysis.batch_analysis?.common_praise || []).map(p => `
  • ${p}
    • `).join('') || '
  • No significant praise detected
    • '}
\n
\n\n
\n

Common Complaints

\n
    ${(aiAnalysis.batch_analysis?.common_complaints || []).map(c => `
  • ${c}
    • `).join('') || '
  • No major complaints found
    • '}
\n
\n\n ${aiAnalysis.batch_analysis?.critical_issues?.length ? `\n
\n

Critical Issues (Immediate Attention)

\n ${aiAnalysis.batch_analysis.critical_issues.map(issue => `\n
\n ${issue.issue}\n ${issue.severity.toUpperCase()}\n \u2014 ${issue.mentions} users mentioned\n
\n `).join('')}\n
` : ''}\n\n
\n

Recommended Actions

\n
    \n
  • Product Team: ${aiAnalysis.recommended_actions?.product_team || 'N/A'}
    • \n
  • Sales Team: ${aiAnalysis.recommended_actions?.sales_team || 'N/A'}
    • \n
  • Support Team: ${aiAnalysis.recommended_actions?.support_team || 'N/A'}
    • \n
\n
\n
\n\n
\n Generated automatically by Thordata Review Intelligence Engine
\n ${new Date().toLocaleString('en-US', { dateStyle: 'long', timeStyle: 'medium' })}\n
\n
\n\n`;\n\nreturn [{\n json: {\n email_html: html,\n product_name: product,\n total_reviews: totalReviews,\n avg_rating: avgRating,\n ai_analysis: aiAnalysis\n }\n}];" }, "typeVersion": 2 }, { "id": "202ad2fa-de6d-491b-85a3-6fe7d044af9e", "name": "Send Executive Summary Email", "type": "n8n-nodes-base.gmail", "position": [ 1152, 3424 ], "parameters": { "sendTo": "user@example.com", "message": "=={{ $json.email_html }}", "options": { "appendAttribution": false }, "subject": "={{ $json.product_name }} Review Analysis \u2013 {{ new Date().toLocaleDateString() }}" }, "credentials": { "gmailOAuth2": { "name": "" } }, "typeVersion": 2.1 }, { "id": "30be5cf7-662f-4c0f-94ec-0c48a460ebef", "name": "Sticky Note", "type": "n8n-nodes-base.stickyNote", "position": [ -208, 2160 ], "parameters": { "width": 1088, "height": 864, "content": "## \ud83c\udf1f Scrape & AI-Analyze Reviews from 5 Platforms in One Click\n\nAutomatically collect hundreds of customer reviews from Trustpilot, Capterra, Chrome Web Store, TrustRadius, and Product Hunt \u2192 deduplicate \u2192 let GPT-4.1 analyze sentiment, praises, complaints, churn risk, and give actionable recommendations \u2192 receive a beautiful executive email report.\n\n### Who\u2019s it for \n- Product managers & founders \n- Growth/marketing teams \n- Customer success & support leads \n- Agencies delivering review reports \n\n### How it works \n1. Submit product URLs (form, webhook or defaults) \n2. Smart pagination + Cloudflare-safe scraping \n3. Universal parser standardizes all reviews \n4. AI collective analysis (not review-by-review) \n5. Gorgeous HTML summary emailed automatically \n\n### Requirements \n- Thordata API key (free tier OK) \u2192 HTTP Header Auth credential \n- OpenAI API key \n- Gmail (or swap for any email node) \n\n### How to set up \n1. Add your Thordata key \u2192 create HTTP Header Auth credential \n2. Add OpenAI credential \n3. Connect Gmail \n4. Test instantly with defaults (Thordata reviews) \n\n### Customize easily \n- Change default product in \u201cPrepare Review Sources\u201d \n- Edit AI prompt for different analysis style \n- Tweak the email design in \u201cRender Elegant Email Template\u201d \n\nPlug-and-play \u00b7 No browser automation \u00b7 Fully deduplicated \u00b7 Rate-limit safe" }, "typeVersion": 1 }, { "id": "24334689-b603-4ddd-8a40-4083b3852cc5", "name": "Sticky Note1", "type": "n8n-nodes-base.stickyNote", "position": [ -1136, 3616 ], "parameters": { "color": 7, "width": 368, "height": 128, "content": "### \ud83d\udd04 STEP 1: Choose your input \n\u2022 Manual Trigger \u2192 quick test with defaults \n\u2022 Form Trigger \u2192 user-friendly UI \n\u2022 Webhook \u2192 integrate with Zapier/Make/etc." }, "typeVersion": 1 }, { "id": "42c3e5a6-e35c-446a-b52a-a94eaa95a3ef", "name": "Sticky Note2", "type": "n8n-nodes-base.stickyNote", "position": [ -416, 3904 ], "parameters": { "color": 7, "width": 896, "height": 96, "content": "### \ud83d\udee0\ufe0f STEP 2: Build source list \nAccepts Trustpilot domain, full URLs for Capterra/Chrome/TrustRadius, and Product Hunt slug \u2192 creates clean source objects." }, "typeVersion": 1 }, { "id": "2de7f216-99a1-4519-864b-dbd102181e89", "name": "Sticky Note3", "type": "n8n-nodes-base.stickyNote", "position": [ -416, 3360 ], "parameters": { "color": 7, "width": 416, "content": "### \ud83d\udd04 STEP 3: Smart pagination loop (rate-limit safe) \n\u2022 Handles all 5 sites differently \n\u2022 Builds correct ?page= URLs \n\u2022 Scrapes via Thordata (JS-rendered + proxy rotation) \n\u2022 Stops when no more reviews on page \n\u2022 2-second polite delay between requests" }, "typeVersion": 1 }, { "id": "7b3943e3-dbf4-4fb4-8073-68ed2f079ff2", "name": "Sticky Note4", "type": "n8n-nodes-base.stickyNote", "position": [ 80, 3216 ], "parameters": { "color": 7, "width": 352, "height": 128, "content": "### \ud83e\uddf9 STEP 4: Global storage + deduplication \nCollects reviews from every source & page \u2192 unique IDs prevent duplicates even if formatting changes." }, "typeVersion": 1 }, { "id": "7653a681-20bc-4680-ac30-d7524d95f330", "name": "Sticky Note5", "type": "n8n-nodes-base.stickyNote", "position": [ 912, 3232 ], "parameters": { "color": 5, "width": 416, "height": 128, "content": "### \ud83e\udd16 STEP 5: AI analysis & beautiful report \n\u2022 All reviews sent together to GPT-4.1 \n\u2022 Structured insights + sentiment scoring \n\u2022 Renders responsive HTML email \n\u2022 Sends executive summary automatically" }, "typeVersion": 1 } ], "connections": { "URL Builder": { "main": [ [ { "node": "Scrape Page (Robust) via Thordata API", "type": "main", "index": 0 } ] ] }, "Manual Trigger": { "main": [ [ { "node": "Prepare Review Sources", "type": "main", "index": 0 } ] ] }, "Has More Pages?": { "main": [ [ { "node": "Increment Page Counter", "type": "main", "index": 0 } ], [ { "node": "Loop Through Sources (Rate Limit Safe)", "type": "main", "index": 0 } ] ] }, "Webhook Trigger": { "main": [ [ { "node": "Prepare Review Sources", "type": "main", "index": 0 } ] ] }, "Initialize Pagination": { "main": [ [ { "node": "URL Builder", "type": "main", "index": 0 } ] ] }, "Rate Limit Delay (2s)": { "main": [ [ { "node": "URL Builder", "type": "main", "index": 0 } ] ] }, "Universal Safe Parser": { "main": [ [ { "node": "Has More Pages?", "type": "main", "index": 0 } ] ] }, "Increment Page Counter": { "main": [ [ { "node": "Rate Limit Delay (2s)", "type": "main", "index": 0 } ] ] }, "Prepare Review Sources": { "main": [ [ { "node": "Loop Through Sources (Rate Limit Safe)", "type": "main", "index": 0 } ] ] }, "AI Review Sentiment Analysis": { "main": [ [ { "node": "Render Elegant Email Template", "type": "main", "index": 0 } ] ] }, "Prepare Data for AI Analysis": { "main": [ [ { "node": "Return Final Results", "type": "main", "index": 0 }, { "node": "AI Review Sentiment Analysis", "type": "main", "index": 0 } ] ] }, "Send Executive Summary Email": { "main": [ [] ] }, "Render Elegant Email Template": { "main": [ [ { "node": "Send Executive Summary Email", "type": "main", "index": 0 } ] ] }, "Form Trigger \u2013 Submit Sources": { "main": [ [ { "node": "Prepare Review Sources", "type": "main", "index": 0 } ] ] }, "Deduplicate & Summarize All Reviews": { "main": [ [ { "node": "Prepare Data for AI Analysis", "type": "main", "index": 0 } ] ] }, "Scrape Page (Robust) via Thordata API": { "main": [ [ { "node": "Universal Safe Parser", "type": "main", "index": 0 } ] ] }, "Loop Through Sources (Rate Limit Safe)": { "main": [ [ { "node": "Deduplicate & Summarize All Reviews", "type": "main", "index": 0 } ], [ { "node": "Initialize Pagination", "type": "main", "index": 0 } ] ] } } }