{ "name": "Monitor RSS feeds for brand and regulatory mentions with rule-based scoring and email digests", "tags": [], "nodes": [ { "id": "5939f65f3a7bc5e3a499b47e908f5d3e3982", "name": "Sticky Overview", "type": "n8n-nodes-base.stickyNote", "position": [ 0, 64 ], "parameters": { "width": 780, "height": 416, "content": "## Monitor RSS feeds for brand and regulatory mentions\n\nRule-based media monitoring: deterministic JavaScript, no per-run API costs, one SMTP credential. Built for comms teams that scan the news for client departments, PR and brand watchers, and competitor or industry trackers.\n\n### How it works\n1. An hourly trigger starts the run; every setting lives in **Config**.\n2. Feeds are fetched, then articles are matched against your topic rules, scored for relevance and sentiment, entity-tagged, and de-duplicated across runs.\n3. One scored HTML digest per audience: the full digest plus one per route in digest.routes. Empty runs are skipped.\n\n### Setup\n1. Edit **Config**: feeds, topics, entities, recipients.\n2. Attach an SMTP credential on **Send Email**.\n3. Run once to test, then activate.\n\nThe notes on each section explain the details." }, "typeVersion": 1 }, { "id": "c4f0662f3531872ed1f5937cdf685b39e835", "name": "Sticky Section A", "type": "n8n-nodes-base.stickyNote", "position": [ 0, 528 ], "parameters": { "color": 7, "width": 452, "height": 360, "content": "### 1. Schedule + Config\nHourly by default; change the interval in the trigger. **Config** is the only node you edit: feeds, topics, entities, scoring, lexicon, and the digest recipients and routes." }, "typeVersion": 1 }, { "id": "46c5b947af7f328743e1f82dd6079ac42ba2", "name": "Sticky Section B", "type": "n8n-nodes-base.stickyNote", "position": [ 480, 528 ], "parameters": { "color": 7, "width": 732, "height": 360, "content": "### 2. Fetch + score\nOne item per feed; a broken feed does not stop the rest. **Process Articles** applies whole-word topic rules, scores relevance and sentiment, tags entities, and de-duplicates across runs via workflow static data." }, "typeVersion": 1 }, { "id": "80ddd5550df8e3cab4380a7c90c30324e1b9", "name": "Sticky Section C", "type": "n8n-nodes-base.stickyNote", "position": [ 1248, 528 ], "parameters": { "color": 7, "width": 684, "height": 508, "content": "### 3. Digest + send\n**Build Digest** renders one scored HTML email per audience: the full digest plus one per route (per client department). **Has Matches?** skips empty runs unless digest.sendEmpty is true. Recipient addresses are placeholders in Config." }, "typeVersion": 1 }, { "id": "128b5176ada25c083207866cba61c71c7e93", "name": "Sticky SMTP Credential", "type": "n8n-nodes-base.stickyNote", "position": [ 1648, 640 ], "parameters": { "color": 3, "width": 236, "height": 92, "content": "### Required\nAttach an SMTP credential on **Send Email** before the first run." }, "typeVersion": 1 }, { "id": "98bbe02e692aac33394b557ce4966bb1ff32", "name": "Schedule Trigger", "type": "n8n-nodes-base.scheduleTrigger", "position": [ 48, 720 ], "parameters": { "rule": { "interval": [ { "field": "hours", "hoursInterval": 1 } ] } }, "typeVersion": 1.2 }, { "id": "43ee64b6538a9b2329089f07db254a64eb65", "name": "Config", "type": "n8n-nodes-base.code", "notes": "Edit feeds[], topics[], entities, lexicon, scoring, digest, seenCap.\nSee README \u00a7Configure and examples/config.example.js.", "position": [ 288, 720 ], "parameters": { "mode": "runOnceForAllItems", "jsCode": "return [{\n json: {\n // ---- 1. Sources ----------------------------------------------------------\n // Any RSS/Atom URL. One bad feed will not stop the others (RSS Read is set\n // to continueRegularOutput).\n feeds: [\n \"https://www.theverge.com/rss/index.xml\",\n \"https://techcrunch.com/feed/\",\n \"https://www.reuters.com/business/finance/rss\",\n \"https://feeds.bbci.co.uk/news/business/rss.xml\"\n ],\n\n // ---- 2. Topics -----------------------------------------------------------\n // Each topic has an include list (ALL terms must be present, boolean AND)\n // and an exclude list (any match \u2192 topic skipped). Whole-word, case-\n // insensitive. An article that matches NO topic is dropped.\n topics: [\n {\n name: \"BrandMentions\",\n include: [\"acme\"],\n exclude: [\"acme tool company\", \"wile e coyote\"]\n },\n {\n name: \"Competitors\",\n include: [\"globex\"],\n exclude: []\n },\n {\n name: \"RegulatoryNews\",\n include: [\"regulation\"],\n exclude: [\"sports\", \"fashion\"]\n }\n ],\n\n // ---- 3. Scoring ----------------------------------------------------------\n // Relevance score = clamp(0..100, round(\n // termWeight * min(totalIncludeHits, 10)\n // + sourceWeight * min(sources[host] || sources.default || 1, 2)\n // + recencyWeight * max(0, 1 - hoursOld / recencyHalfLifeHours)\n // ))\n scoring: {\n termWeight: 6, // raise to ~12 if you want keyword density to dominate (good for very specific brand names)\n sourceWeight: 20, // raise if you trust a few outlets a lot more than the rest\n recencyWeight: 30, // lower to ~10 for slow-moving regulatory/industry feeds; raise for breaking-news monitoring\n recencyHalfLifeHours: 48, // raise to 168 (1 week) for weekly digests; lower to 12 for \"what happened overnight\"\n sources: {\n \"reuters.com\": 1.5,\n \"bbc.co.uk\": 1.4,\n \"techcrunch.com\": 1.2,\n \"theverge.com\": 1.1,\n default: 1.0 // anything not listed above gets this multiplier\n }\n },\n\n // ---- 4. Sentiment lexicon -----------------------------------------------\n // AFINN-style word \u2192 integer score (typical \u22125..+5). Sum is normalized by\n // sqrt(tokenCount) so long articles don't dominate. Label thresholds:\n // \u2265 +0.5 \u2192 positive, \u2264 -0.5 \u2192 negative, else neutral.\n lexicon: {\n surge: 3, soar: 3, beat: 2, growth: 2, profit: 2, win: 2, gain: 2,\n record: 1, upbeat: 2, strong: 1, rise: 1, rally: 2,\n loss: -2, miss: -2, slump: -3, plunge: -3, fall: -1, drop: -1,\n lawsuit: -3, fraud: -4, breach: -3, scandal: -3, fine: -2, probe: -2,\n layoff: -3, layoffs: -3, recall: -2, hack: -3, outage: -2\n },\n\n // ---- 5. Entity dictionary -----------------------------------------------\n // Label \u2192 list of aliases. Whole-word, case-insensitive. Each label is\n // attached at most once per article.\n entities: {\n \"Acme Corp\": [\"acme\", \"acme corp\", \"acmecorp\"],\n \"Globex Inc\": [\"globex\", \"globex corp\"],\n \"Initech\": [\"initech\"]\n },\n\n // ---- 6. Digest email ----------------------------------------------------\n digest: {\n subjectPrefix: \"[MediaMonitor]\",\n minRelevance: 30, // hide low-score noise from the email\n maxItems: 50, // hard cap across all topics\n to: \"ops@example.com\", // full-digest recipient; set to \"\" to send routed digests only\n from: \"monitor@example.com\",\n sendEmpty: false, // true = also email runs with zero matches (\"No new matches\")\n\n // Routes: one EXTRA email per entry, filtered to the listed topics.\n // Lets a comms desk send each client department only its own coverage.\n // An article matching two routed topics appears in both emails.\n routes: [\n // { name: \"Environment Dept\", to: \"env-comms@example.gov\", topics: [\"RegulatoryNews\"] },\n // { name: \"Trade Dept\", to: \"trade-comms@example.gov\", topics: [\"Competitors\"] }\n ]\n },\n\n // ---- 7. Dedup --------------------------------------------------------\n // Max number of link hashes kept in workflow static data. Older entries\n // are dropped FIFO.\n seenCap: 5000\n }\n}];\n\n// ---- Starter-pack topics (uncomment and rename to use) ----------------------\n// Drop any of these into topics[] above. Replace the placeholder terms with\n// your real brand/competitor/regulator/industry keywords.\n//\n// { name: \"OwnBrand\", include: [\"yourbrand\"], exclude: [] },\n// { name: \"Top3Competitors\", include: [\"competitor\"], exclude: [] },\n// { name: \"RegulatoryRisk\", include: [\"regulator\"], exclude: [\"sports\"] },\n// { name: \"IndustryGeneral\", include: [\"fintech\"], exclude: [\"crypto\"] },\n// { name: \"Tourism\", include: [\"tourism\"], exclude: [] },\n// { name: \"Hospitality\", include: [\"hotel\"], exclude: [\"hotel california\"] },\n// { name: \"Cruise\", include: [\"cruise\"], exclude: [\"tom cruise\", \"cruise missile\"] },\n", "language": "javaScript" }, "notesInFlow": true, "typeVersion": 2 }, { "id": "c4d445bc0b739664687dc54c6e23d3fa308c", "name": "Feed URLs", "type": "n8n-nodes-base.code", "position": [ 528, 720 ], "parameters": { "mode": "runOnceForAllItems", "jsCode": "// Fan out: one item per feed URL from Config.\nconst feeds = $('Config').first().json.feeds || [];\nreturn feeds.map(url => ({ json: { url } }));\n", "language": "javaScript" }, "typeVersion": 2 }, { "id": "33019dc91262e668ad526153e9d04d8a149c", "name": "RSS Read", "type": "n8n-nodes-base.rssFeedRead", "onError": "continueRegularOutput", "position": [ 768, 720 ], "parameters": { "url": "={{ $json.url }}", "options": {} }, "typeVersion": 1 }, { "id": "a8391248c70e77f060b11f71d5a3ec327df1", "name": "Process Articles", "type": "n8n-nodes-base.code", "notes": "The brain. Rule-based enrichment: topic match, relevance, sentiment, entities, dedup.\nPure functions inlined from src/lib.mjs. Reads $(\"Config\").first().json + workflow static data.", "position": [ 1008, 720 ], "parameters": { "mode": "runOnceForAllItems", "jsCode": "// Process Articles: rule-based enrichment.\n// Pure functions are inlined below (copy of src/lib.mjs) so this node is\n// self-contained inside the exported workflow. Edit src/lib.mjs and re-run\n// scripts/build-workflow.mjs to regenerate.\n\nconst HTML_ENTITIES = {\n '&': '&', '<': '<', '>': '>',\n '"': '\"', ''': \"'\", ''': \"'\", ' ': ' '\n};\n\nfunction stripHtml(s) {\n if (!s) return '';\n return String(s)\n .replace(/<\\s*br\\s*\\/?\\s*>/gi, ' ')\n .replace(/<\\/?\\s*(p|div|li|h[1-6])\\s*>/gi, ' ')\n .replace(/<[^>]+>/g, '')\n .replace(/&(amp|lt|gt|quot|#39|apos|nbsp);/g, m => HTML_ENTITIES[m] ?? m)\n .replace(/\\s+/g, ' ')\n .trim();\n}\n\nfunction safeHostname(link) {\n try {\n return new URL(link).hostname.replace(/^www\\./, '').toLowerCase();\n } catch {\n return '';\n }\n}\n\nfunction escapeRegex(s) {\n return String(s).replace(/[.*+?^${}()|[\\]\\\\]/g, '\\\\$&');\n}\n\nfunction termRegex(term) {\n // Whole-word, case-insensitive. \\b works for ASCII; phrases with spaces\n // get \\b on the outer tokens which is the common-case intent.\n return new RegExp(`\\\\b${escapeRegex(term)}\\\\b`, 'gi');\n}\n\nfunction countMatches(haystack, term) {\n const m = haystack.match(termRegex(term));\n return m ? m.length : 0;\n}\n\nfunction hasMatch(haystack, term) {\n return termRegex(term).test(haystack);\n}\n\n/**\n * Cut a summary at `n` characters on a word boundary and append an ellipsis.\n * Used by the digest builder for tidy article previews.\n * @param {string} text - input string\n * @param {number} [n=280] - max length before truncation\n * @returns {string} truncated text, or original if already short enough\n */\nfunction truncateSummary(text, n = 280) {\n const s = String(text ?? '');\n if (s.length <= n) return s;\n const cut = s.slice(0, n);\n const lastSpace = cut.lastIndexOf(' ');\n return (lastSpace > n * 0.6 ? cut.slice(0, lastSpace) : cut) + '\u2026';\n}\n\n/**\n * Normalize a raw RSS/Atom item into a flat article shape.\n * Strips HTML, picks the best content field, derives `source` from the link host,\n * and parses `publishedAt` to ISO 8601.\n * @param {object} raw - raw feed item (from n8n RSS Read node)\n * @param {string} feedUrl - the feed URL (fallback for source host)\n * @returns {{title:string, link:string, source:string, publishedAt:string, summary:string, contentText:string}}\n */\nfunction normalizeArticle(raw, feedUrl) {\n const title = stripHtml(raw?.title ?? '');\n const link = String(raw?.link ?? raw?.url ?? raw?.guid ?? '').trim();\n const summary = stripHtml(raw?.contentSnippet ?? raw?.summary ?? raw?.description ?? '');\n const content = stripHtml(raw?.content ?? raw?.['content:encoded'] ?? '');\n const contentText = (content && content.length > summary.length) ? content : summary;\n const published = raw?.isoDate ?? raw?.pubDate ?? raw?.published ?? raw?.date ?? null;\n let publishedAt;\n if (published) {\n const d = new Date(published);\n publishedAt = isNaN(d.getTime()) ? new Date().toISOString() : d.toISOString();\n } else {\n publishedAt = new Date().toISOString();\n }\n const source = safeHostname(link) || safeHostname(feedUrl) || '';\n return { title, link, source, publishedAt, summary, contentText };\n}\n\n/**\n * Hash a URL to a short stable fingerprint for cross-run dedup.\n * Lowercases, strips tracking query params (utm_*, gclid, fbclid, mc_cid, mc_eid),\n * drops trailing slash and fragment, then djb2 \u2192 32-bit hex.\n * @param {string} link\n * @returns {string} 8-char hex fingerprint, '0' for empty input\n */\nfunction hashLink(link) {\n if (!link) return '0';\n let normalized = String(link).trim().toLowerCase();\n try {\n const u = new URL(normalized);\n u.hash = '';\n const drop = [];\n for (const k of u.searchParams.keys()) {\n if (k.startsWith('utm_') || k === 'gclid' || k === 'fbclid' || k === 'mc_cid' || k === 'mc_eid') {\n drop.push(k);\n }\n }\n drop.forEach(k => u.searchParams.delete(k));\n normalized = `${u.protocol}//${u.hostname}${u.pathname.replace(/\\/$/, '')}${u.search}`;\n } catch {\n normalized = normalized.replace(/#.*$/, '').replace(/\\/$/, '');\n }\n // djb2 32-bit\n let h = 5381;\n for (let i = 0; i < normalized.length; i++) {\n h = (((h << 5) + h) + normalized.charCodeAt(i)) | 0;\n }\n return (h >>> 0).toString(16);\n}\n\n/**\n * Whole-word, case-insensitive boolean topic match.\n * For each topic: ALL include[] terms must appear AND no exclude[] term may appear.\n * @param {{title?:string, contentText?:string}} article\n * @param {Array<{name:string, include:string[], exclude:string[]}>} topics\n * @returns {string[]} names of topics matched, [] if none\n */\nfunction matchTopics(article, topics) {\n const hay = `${article.title ?? ''} ${article.contentText ?? ''}`;\n const matched = [];\n for (const t of topics ?? []) {\n const include = t.include ?? [];\n const exclude = t.exclude ?? [];\n if (include.length === 0) continue;\n const allInclude = include.every(term => hasMatch(hay, term));\n if (!allInclude) continue;\n const anyExclude = exclude.some(term => hasMatch(hay, term));\n if (anyExclude) continue;\n matched.push(t.name);\n }\n return matched;\n}\n\n/**\n * Compute a relevance score 0\u2013100 from term frequency, source weight, recency.\n * @param {{title?:string, contentText?:string, source?:string, publishedAt?:string}} article\n * @param {string[]} topicsMatched - names of topics this article matched\n * @param {object} scoring - { termWeight, sourceWeight, recencyWeight, recencyHalfLifeHours, sources }\n * @param {Array<{name:string, include:string[]}>} allTopics\n * @param {number} [now=Date.now()] - reference timestamp for recency math\n * @returns {number} integer 0..100\n */\nfunction scoreRelevance(article, topicsMatched, scoring, allTopics, now = Date.now()) {\n if (!topicsMatched || topicsMatched.length === 0) return 0;\n const cfg = scoring ?? {};\n const termWeight = cfg.termWeight ?? 6;\n const sourceWeight = cfg.sourceWeight ?? 20;\n const recencyWeight = cfg.recencyWeight ?? 30;\n const halfLife = cfg.recencyHalfLifeHours ?? 48;\n const sources = cfg.sources ?? {};\n\n const hay = `${article.title ?? ''} ${article.contentText ?? ''}`;\n let hits = 0;\n for (const tName of topicsMatched) {\n const t = (allTopics ?? []).find(x => x.name === tName);\n if (!t) continue;\n for (const term of t.include ?? []) {\n hits += countMatches(hay, term);\n }\n }\n // cap hits so a keyword-stuffed article doesn't dominate\n const cappedHits = Math.min(hits, 10);\n const termComponent = cappedHits * termWeight;\n\n const srcMultiplier = sources[article.source] ?? sources.default ?? 1;\n const sourceComponent = sourceWeight * Math.min(srcMultiplier, 2); // clamp absurd weights\n\n const published = new Date(article.publishedAt ?? now).getTime();\n const hoursOld = Math.max(0, (now - published) / 36e5);\n const recencyComponent = recencyWeight * Math.max(0, 1 - hoursOld / halfLife);\n\n const total = termComponent + sourceComponent + recencyComponent;\n return Math.max(0, Math.min(100, Math.round(total)));\n}\n\n/**\n * AFINN-style sentiment. Sum lexicon[word] across whole-word tokens, normalize\n * by sqrt(tokenCount) so long articles don't dominate.\n * @param {{title?:string, contentText?:string}} article\n * @param {Record} lexicon - word \u2192 integer (typical \u22125..+5)\n * @returns {{score:number, label:'positive'|'neutral'|'negative'}}\n */\nfunction scoreSentiment(article, lexicon) {\n const text = `${article.title ?? ''} ${article.contentText ?? ''}`.toLowerCase();\n const tokens = text.match(/[a-z][a-z'-]+/g) ?? [];\n if (tokens.length === 0) return { score: 0, label: 'neutral' };\n let raw = 0;\n for (const tok of tokens) {\n const v = lexicon?.[tok];\n if (typeof v === 'number') raw += v;\n }\n const score = raw / Math.sqrt(tokens.length);\n const rounded = Math.round(score * 100) / 100;\n let label = 'neutral';\n if (rounded >= 0.5) label = 'positive';\n else if (rounded <= -0.5) label = 'negative';\n return { score: rounded, label };\n}\n\n/**\n * Tag entities by alias. Each label is attached at most once per article.\n * @param {{title?:string, contentText?:string}} article\n * @param {Record} entities - label \u2192 aliases (whole-word, case-insensitive)\n * @returns {string[]} unique labels matched\n */\nfunction tagEntities(article, entities) {\n const hay = `${article.title ?? ''} ${article.contentText ?? ''}`;\n const found = new Set();\n for (const [label, aliases] of Object.entries(entities ?? {})) {\n for (const alias of aliases ?? []) {\n if (hasMatch(hay, alias)) { found.add(label); break; }\n }\n }\n return [...found];\n}\n\n\nconst cfg = $('Config').first().json;\nconst store = $getWorkflowStaticData('global');\nif (!Array.isArray(store.seen)) store.seen = [];\nconst seenSet = new Set(store.seen);\n\nconst now = Date.now();\nconst enriched = [];\nconst newHashes = [];\n\nfor (const item of items) {\n const raw = item.json || {};\n const article = normalizeArticle(raw, raw.feedUrl || '');\n if (!article.link) continue;\n\n const topicsMatched = matchTopics(article, cfg.topics);\n if (topicsMatched.length === 0) continue;\n\n const hash = hashLink(article.link);\n if (seenSet.has(hash)) continue;\n seenSet.add(hash);\n newHashes.push(hash);\n\n const relevance = scoreRelevance(article, topicsMatched, cfg.scoring, cfg.topics, now);\n const sentiment = scoreSentiment(article, cfg.lexicon);\n const entitiesFound = tagEntities(article, cfg.entities);\n\n enriched.push({\n title: article.title,\n link: article.link,\n source: article.source,\n publishedAt: article.publishedAt,\n summary: article.summary,\n topics: topicsMatched.join(', '),\n topicsList: topicsMatched,\n entities: entitiesFound.join(', '),\n entitiesList: entitiesFound,\n relevance,\n sentiment: sentiment.label,\n sentimentScore: sentiment.score,\n hash,\n scannedAt: new Date(now).toISOString()\n });\n}\n\n// Persist seen-list, trimmed FIFO.\nstore.seen = [...store.seen, ...newHashes].slice(-1 * (cfg.seenCap || 5000));\n\nenriched.sort((a, b) => b.relevance - a.relevance);\nreturn enriched.map(a => ({ json: a }));\n", "language": "javaScript" }, "notesInFlow": true, "typeVersion": 2 }, { "id": "bb6d47b5e1b1c7c47a02b991efae0ff4aa46", "name": "Build Digest", "type": "n8n-nodes-base.code", "position": [ 1264, 720 ], "parameters": { "mode": "runOnceForAllItems", "jsCode": "// Build Digest: render scored HTML email digests, one output item per email.\n// Default: a single full digest to digest.to. With digest.routes configured,\n// also one filtered digest per route so each team gets only its own topics.\n// Every output item carries a 'total' count used by Has Matches? to skip empty runs.\nconst cfg = $('Config').first().json;\nconst digestCfg = cfg.digest || {};\nconst minRel = digestCfg.minRelevance ?? 0;\nconst maxItems = digestCfg.maxItems ?? 50;\n\nfunction escapeHtml(s) {\n return String(s ?? '')\n .replace(/&/g, '&').replace(//g, '>')\n .replace(/\"/g, '"').replace(/'/g, ''');\n}\n\nfunction badge(label, color) {\n return `${escapeHtml(label)}`;\n}\n\nfunction sentimentColor(s) {\n if (s === 'positive') return '#137333';\n if (s === 'negative') return '#b3261e';\n return '#5f6368';\n}\nfunction relevanceColor(r) {\n if (r >= 70) return '#0b8043';\n if (r >= 40) return '#ef6c00';\n return '#9aa0a6';\n}\n\nconst all = items\n .map(i => i.json)\n .filter(a => (a.relevance ?? 0) >= minRel)\n .slice(0, maxItems);\n\nconst dateStr = new Date().toLocaleString('en-US', { dateStyle: 'medium', timeStyle: 'short' });\n\n// route = null renders the full digest; a route object filters to route.topics.\nfunction renderDigest(articles, route) {\n const topicFilter = route && Array.isArray(route.topics) && route.topics.length ? route.topics : null;\n const byTopic = {};\n const counted = new Set();\n for (const a of articles) {\n for (const t of (a.topicsList || [])) {\n if (topicFilter && !topicFilter.includes(t)) continue;\n (byTopic[t] = byTopic[t] || []).push(a);\n counted.add(a.hash || a.link);\n }\n }\n const topicNames = Object.keys(byTopic).sort();\n const total = counted.size;\n const audience = route && route.name ? route.name + ': ' : '';\n const subject = `${digestCfg.subjectPrefix || '[MediaMonitor]'} ${audience}${total} match${total === 1 ? '' : 'es'}: ${dateStr}`;\n\n let body = `\n
\n

Media Monitor digest${route && route.name ? ' · ' + escapeHtml(route.name) : ''}

\n
\n ${dateStr} · ${total} new match${total === 1 ? '' : 'es'}\n ${topicNames.length ? '· topics: ' + topicNames.map(escapeHtml).join(', ') : ''}\n
\n`;\n\n if (total === 0) {\n body += '

No new matches in this run.

';\n } else {\n for (const topic of topicNames) {\n const articles2 = byTopic[topic].slice().sort((a, b) => b.relevance - a.relevance);\n body += `

${escapeHtml(topic)} (${articles2.length})

`;\n body += '
    ';\n for (const a of articles2) {\n const published = a.publishedAt ? new Date(a.publishedAt).toLocaleString('en-US', { dateStyle: 'medium', timeStyle: 'short' }) : '';\n const ents = (a.entitiesList || []).map(e => badge(e, '#1a73e8')).join('');\n body += `\n
  • \n
    \n ${badge(a.relevance + '%', relevanceColor(a.relevance))}\n ${badge(a.sentiment, sentimentColor(a.sentiment))}\n ${escapeHtml(a.source)} · ${escapeHtml(published)}\n
    \n \n
    \n ${escapeHtml((a.summary || '').slice(0, 280))}${(a.summary || '').length > 280 ? '…' : ''}\n
    \n ${ents ? '
    ' + ents + '
    ' : ''}\n
    • `;\n }\n body += '
';\n }\n }\n\n body += `\n
\n
media-monitor · rule-based · self-hosted n8n
\n
`;\n\n return { subject, html: body, total };\n}\n\nconst out = [];\n\nif (digestCfg.to) {\n const full = renderDigest(all, null);\n out.push({ json: { to: digestCfg.to, from: digestCfg.from, subject: full.subject, html: full.html, total: full.total } });\n}\n\nfor (const route of (Array.isArray(digestCfg.routes) ? digestCfg.routes : [])) {\n if (!route || !route.to || !Array.isArray(route.topics) || route.topics.length === 0) continue;\n const r = renderDigest(all, route);\n out.push({ json: { to: route.to, from: digestCfg.from, subject: r.subject, html: r.html, total: r.total } });\n}\n\nreturn out;\n", "language": "javaScript" }, "typeVersion": 2 }, { "id": "5f403372505f33b33530a994c2f1286c0282", "name": "Has Matches?", "type": "n8n-nodes-base.if", "notes": "Drops digests with zero matches unless digest.sendEmpty is true.", "position": [ 1504, 720 ], "parameters": { "options": {}, "conditions": { "options": { "version": 2, "leftValue": "", "caseSensitive": true, "typeValidation": "loose" }, "combinator": "and", "conditions": [ { "id": "has-matches-or-send-empty", "operator": { "type": "boolean", "operation": "true", "singleValue": true }, "leftValue": "={{ $json.total > 0 || $('Config').first().json.digest.sendEmpty === true }}", "rightValue": "" } ] } }, "notesInFlow": true, "typeVersion": 2.2 }, { "id": "7d607f4d04463ae4b7f3dded9d9c10ce4afd", "name": "Send Email", "type": "n8n-nodes-base.emailSend", "position": [ 1744, 752 ], "parameters": { "html": "={{ $json.html }}", "options": {}, "subject": "={{ $json.subject }}", "toEmail": "={{ $json.to }}", "fromEmail": "={{ $json.from }}", "emailFormat": "html" }, "typeVersion": 2 }, { "id": "f840b6939e72f4f6eae9845ae7eaca74a217", "name": "Skip Empty Run", "type": "n8n-nodes-base.noOp", "position": [ 1744, 896 ], "parameters": {}, "typeVersion": 1 } ], "active": false, "settings": { "executionOrder": "v1" }, "versionId": "195887d29e34cd17a20ae6e8e2061305cc31", "connections": { "Config": { "main": [ [ { "node": "Feed URLs", "type": "main", "index": 0 } ] ] }, "RSS Read": { "main": [ [ { "node": "Process Articles", "type": "main", "index": 0 } ] ] }, "Feed URLs": { "main": [ [ { "node": "RSS Read", "type": "main", "index": 0 } ] ] }, "Build Digest": { "main": [ [ { "node": "Has Matches?", "type": "main", "index": 0 } ] ] }, "Has Matches?": { "main": [ [ { "node": "Send Email", "type": "main", "index": 0 } ], [ { "node": "Skip Empty Run", "type": "main", "index": 0 } ] ] }, "Process Articles": { "main": [ [ { "node": "Build Digest", "type": "main", "index": 0 } ] ] }, "Schedule Trigger": { "main": [ [ { "node": "Config", "type": "main", "index": 0 } ] ] } } }