{ "name": "Blog - YouTube Analyze", "nodes": [ { "parameters": { "httpMethod": "POST", "path": "blog-youtube-analyze", "responseMode": "responseNode", "options": {} }, "id": "webhook-trigger", "name": "Webhook", "type": "n8n-nodes-base.webhook", "position": [ 200, 300 ], "typeVersion": 2 }, { "parameters": { "jsCode": "// Required env vars: SUPABASE_URL, SUPABASE_SERVICE_KEY, ANTHROPIC_API_KEY\n// Optional: YOUTUBE_API_KEY (enables richer metadata via Data API v3)\n\nconst body = $input.first().json;\nconst url = body.url || '';\nconst videoId = body.video_id || url.match(/(?:v=|youtu\\.be\\/|embed\\/)([a-zA-Z0-9_-]{11})/)?.[1];\n\nif (!videoId) {\n throw new Error('Missing video_id \u2014 could not extract from URL: ' + url);\n}\n\n// 1. Fetch oEmbed metadata (no API key required)\nlet title = '', channelName = '', thumbnailUrl = '';\ntry {\n const oEmbedRes = await fetch(\n `https://www.youtube.com/oembed?url=${encodeURIComponent(url)}&format=json`\n );\n if (oEmbedRes.ok) {\n const oe = await oEmbedRes.json();\n title = oe.title || '';\n channelName = oe.author_name || '';\n thumbnailUrl = `https://i.ytimg.com/vi/${videoId}/hqdefault.jpg`;\n }\n} catch (e) {\n console.error('oEmbed fetch failed:', e.message);\n}\n\n// 2. Try to fetch transcript via YouTube timedtext API (works for many videos)\nlet transcript = '';\ntry {\n const transcriptRes = await fetch(\n `https://www.youtube.com/api/timedtext?lang=en&v=${videoId}`\n );\n if (transcriptRes.ok) {\n const xml = await transcriptRes.text();\n // Strip XML tags, decode entities\n transcript = xml\n .replace(/<[^>]+>/g, ' ')\n .replace(/&/g, '&')\n .replace(/"/g, '\"')\n .replace(/'/g, \"'\")\n .replace(/</g, '<')\n .replace(/>/g, '>')\n .replace(/\\s+/g, ' ')\n .trim()\n .slice(0, 12000); // Claude Haiku context limit\n }\n} catch (e) {\n console.error('Timedtext fetch failed:', e.message);\n}\n\n// 3. If YOUTUBE_API_KEY is set, enrich with Data API v3\nconst ytApiKey = process.env.YOUTUBE_API_KEY;\nlet durationSeconds = null;\nif (ytApiKey) {\n try {\n const apiRes = await fetch(\n `https://www.googleapis.com/youtube/v3/videos?id=${videoId}&part=snippet,contentDetails&key=${ytApiKey}`\n );\n if (apiRes.ok) {\n const apiData = await apiRes.json();\n const item = apiData.items?.[0];\n if (item) {\n title = title || item.snippet?.title || '';\n channelName = channelName || item.snippet?.channelTitle || '';\n // Parse ISO 8601 duration (PT4M13S \u2192 253 seconds)\n const dur = item.contentDetails?.duration || '';\n const m = dur.match(/PT(?:(\\d+)H)?(?:(\\d+)M)?(?:(\\d+)S)?/);\n if (m) {\n durationSeconds = (parseInt(m[1]||'0')*3600) + (parseInt(m[2]||'0')*60) + parseInt(m[3]||'0');\n }\n }\n }\n } catch (e) {\n console.error('YouTube Data API failed:', e.message);\n }\n}\n\n// 4. Insert initial row into Supabase\nconst supabaseUrl = process.env.SUPABASE_URL;\nconst supabaseKey = process.env.SUPABASE_SERVICE_KEY;\n\nif (!supabaseUrl || !supabaseKey) {\n throw new Error('Missing SUPABASE_URL or SUPABASE_SERVICE_KEY env vars');\n}\n\nconst insertRes = await fetch(`${supabaseUrl}/rest/v1/blog_cms_youtube_sources`, {\n method: 'POST',\n headers: {\n 'Content-Type': 'application/json',\n 'apikey': supabaseKey,\n 'Authorization': `Bearer ${supabaseKey}`,\n 'Prefer': 'return=representation'\n },\n body: JSON.stringify({\n url,\n video_id: videoId,\n title,\n channel_name: channelName,\n thumbnail_url: thumbnailUrl,\n duration_seconds: durationSeconds,\n transcript: transcript || null,\n transcript_language: transcript ? 'en' : null,\n status: transcript ? 'transcribed' : 'imported',\n key_topics: [],\n suggested_wiki_terms: [],\n article_opportunities: []\n })\n});\n\nif (!insertRes.ok) {\n const err = await insertRes.text();\n throw new Error(`Supabase insert failed (${insertRes.status}): ${err.slice(0, 300)}`);\n}\n\nconst [insertedRow] = await insertRes.json();\nconst sourceId = insertedRow.id;\n\nreturn [{ json: { source_id: sourceId, video_id: videoId, title, channel_name: channelName, transcript_length: transcript.length, status: 'transcribed' } }];\n" }, "id": "fetch-metadata", "name": "Fetch Metadata + Transcript", "type": "n8n-nodes-base.code", "position": [ 480, 300 ], "typeVersion": 2 }, { "parameters": { "jsCode": "// Analyze transcript with Claude Haiku \u2192 key_topics + article_opportunities\n// Required: ANTHROPIC_API_KEY, SUPABASE_URL, SUPABASE_SERVICE_KEY\n\nconst { source_id, video_id, title, channel_name, transcript_length } = $input.first().json;\n\nconst apiKey = process.env.ANTHROPIC_API_KEY;\nif (!apiKey) throw new Error('Missing ANTHROPIC_API_KEY env var');\n\n// Re-fetch transcript from Supabase\nconst supabaseUrl = process.env.SUPABASE_URL;\nconst supabaseKey = process.env.SUPABASE_SERVICE_KEY;\n\nconst rowRes = await fetch(\n `${supabaseUrl}/rest/v1/blog_cms_youtube_sources?id=eq.${source_id}&select=transcript,title,channel_name`,\n { headers: { 'apikey': supabaseKey, 'Authorization': `Bearer ${supabaseKey}` } }\n);\nconst [row] = await rowRes.json();\nconst transcript = row?.transcript || '';\nconst videoTitle = row?.title || title || '(unknown)';\n\nlet keyTopics = [];\nlet articleOpportunities = [];\nlet contextSummary = '';\n\nif (transcript.length > 100) {\n const prompt = `You are an editorial analyst. Analyze this YouTube transcript and extract structured insights for a blog content team.\n\nVideo: \"${videoTitle}\" by ${channel_name || 'Unknown channel'}\n\nTranscript (first 8000 chars):\n${transcript.slice(0, 8000)}\n\nReturn ONLY valid JSON with this exact structure:\n{\n \"key_topics\": [\"topic1\", \"topic2\", \"topic3\", \"topic4\", \"topic5\"],\n \"article_opportunities\": [\"angle1: description\", \"angle2: description\", \"angle3: description\"],\n \"context_summary\": \"2-3 sentence summary of the main argument and key insights\"\n}\n\nkey_topics: 4-6 short phrases (max 5 words each) covering the main themes.\narticle_opportunities: 3 specific blog angles this content suggests, formatted as \"Angle title: brief description\".\ncontext_summary: What is the core argument and most actionable insight?`;\n\n try {\n const claudeRes = await fetch('https://api.anthropic.com/v1/messages', {\n method: 'POST',\n headers: {\n 'x-api-key': apiKey,\n 'anthropic-version': '2023-06-01',\n 'content-type': 'application/json'\n },\n body: JSON.stringify({\n model: 'claude-haiku-4-5-20251001',\n max_tokens: 800,\n messages: [{ role: 'user', content: prompt }]\n })\n });\n\n if (claudeRes.ok) {\n const claudeData = await claudeRes.json();\n const text = claudeData.content?.[0]?.text || '';\n const jsonMatch = text.match(/\\{[\\s\\S]*\\}/);\n if (jsonMatch) {\n const parsed = JSON.parse(jsonMatch[0]);\n keyTopics = parsed.key_topics || [];\n articleOpportunities = parsed.article_opportunities || [];\n contextSummary = parsed.context_summary || '';\n }\n }\n } catch (e) {\n console.error('Claude analysis failed:', e.message);\n // Continue \u2014 will save with empty topics but mark as analyzed\n }\n} else {\n contextSummary = `No transcript available for \"${videoTitle}\". Topics extracted from title only.`;\n // Extract basic topics from title\n keyTopics = videoTitle\n .split(/[\\s,\\-\u2013:]+/)\n .filter(w => w.length > 4)\n .slice(0, 5);\n}\n\n// Update Supabase row with analysis results\nconst updateRes = await fetch(\n `${supabaseUrl}/rest/v1/blog_cms_youtube_sources?id=eq.${source_id}`,\n {\n method: 'PATCH',\n headers: {\n 'Content-Type': 'application/json',\n 'apikey': supabaseKey,\n 'Authorization': `Bearer ${supabaseKey}`,\n 'Prefer': 'return=minimal'\n },\n body: JSON.stringify({\n key_topics: keyTopics,\n article_opportunities: articleOpportunities,\n context_summary: contextSummary,\n status: 'analyzed',\n updated_at: new Date().toISOString()\n })\n }\n);\n\nif (!updateRes.ok) {\n const err = await updateRes.text();\n console.error(`Supabase update failed (${updateRes.status}): ${err}`);\n}\n\nreturn [{ json: { source_id, key_topics: keyTopics, article_opportunities: articleOpportunities, context_summary: contextSummary, status: 'analyzed' } }];\n" }, "id": "analyze-with-claude", "name": "Analyze with Claude Haiku", "type": "n8n-nodes-base.code", "position": [ 760, 300 ], "typeVersion": 2 }, { "parameters": { "respondWith": "json", "responseBody": "={{ JSON.stringify({ source_id: $json.source_id, status: $json.status, key_topics: $json.key_topics, article_opportunities: $json.article_opportunities }) }}", "options": {} }, "id": "respond-webhook", "name": "Respond to Webhook", "type": "n8n-nodes-base.respondToWebhook", "position": [ 1040, 300 ], "typeVersion": 1 } ], "connections": { "Webhook": { "main": [ [ { "node": "Fetch Metadata + Transcript", "type": "main", "index": 0 } ] ] }, "Fetch Metadata + Transcript": { "main": [ [ { "node": "Analyze with Claude Haiku", "type": "main", "index": 0 } ] ] }, "Analyze with Claude Haiku": { "main": [ [ { "node": "Respond to Webhook", "type": "main", "index": 0 } ] ] } }, "active": false, "settings": { "executionOrder": "v1" }, "meta": { "notes": "Required n8n env vars: SUPABASE_URL, SUPABASE_SERVICE_KEY, ANTHROPIC_API_KEY. Optional: YOUTUBE_API_KEY (enriches metadata via YouTube Data API v3). Activate after verifying env vars are set." } }