mcp-scraper 0.3.28 → 0.3.42

package/dist/{chunk-NGD4ITAO.js → chunk-KPF64WST.js}

@@ -1,6 +1,6 @@
 import {
   PACKAGE_VERSION
-} from "./chunk-7MY2NSQX.js";
+} from "./chunk-OUZZNW3I.js";
 import {
   browserServiceProfileName,
   browserServiceProfileSaveChanges,
@@ -127,6 +127,13 @@ seam is noted so you can chain them.
 - Transcribe an ad video -> **facebook_ad_transcribe** (takes a videoUrl from \`facebook_page_intel\`).
 - Transcribe an organic reel/video/post -> **facebook_video_transcribe** (takes the Facebook url directly).
 
+## Google Ads (Ads Transparency Center)
+- Find advertisers by domain or brand -> **google_ads_search** (returns \`advertisers[]\` with advertiserId).
+- One advertiser's ads -> **google_ads_page_intel** (takes advertiserId or domain; returns \`ads[]\` with
+  format, imageUrls, youtubeVideoId, and videoUrl).
+- Transcribe a video ad -> **google_ads_transcribe** (takes a googlevideo \`videoUrl\` from
+  \`google_ads_page_intel\`), or **youtube_transcribe** for a returned \`youtubeVideoId\`.
+
 ## Instagram
 - Profile inventory -> **instagram_profile_content** (returns post/reel/tv urls).
 - One post or reel -> **instagram_media_download** (takes a url from \`instagram_profile_content\`, or a
@@ -1012,7 +1019,9 @@ ${paaRows}` : "## People Also Ask\n*Google did not return a People Also Ask bloc
 ${serpRows}` : "";
   const aiSection = aiOvw?.detected && aiOvw.text ? `
 ## AI Overview
-> ${truncate(aiOvw.text, 600)}` : "";
+> ${truncate(aiOvw.text, 600)}` + (aiOvw.shareUrl ? `
+
+**Shareable link:** ${aiOvw.shareUrl}` : "") : "";
   const statsLine = durationMs ? `
 ## Stats
 - Status: ${diagnostics?.completionStatus ?? (flat.length ? "paa_found" : "no_paa")} \xB7 Questions: ${flat.length} \xB7 Duration: ${(durationMs / 1e3).toFixed(1)}s` : "";
@@ -1045,7 +1054,7 @@ ${paaTable}${serpTable}${entityIdsSection(entityIds)}${aiSection}${statsLine}${d
         domain: String(r.domain ?? ""),
         snippet: r.snippet ?? null
       })),
-      aiOverview: aiOvw ? { detected: aiOvw.detected === true, text: aiOvw.text ?? null } : null,
+      aiOverview: aiOvw ? { detected: aiOvw.detected === true, text: aiOvw.text ?? null, shareUrl: aiOvw.shareUrl ?? null } : null,
       entityIds: entityIds ? { entities: entityIds.entities ?? [], kgIds: entityIds.kgIds ?? [], cids: entityIds.cids ?? [], gcids: entityIds.gcids ?? [] } : null,
       durationMs: durationMs ?? null
     }
@@ -1077,7 +1086,9 @@ ${serpRows}` : "## Organic Results\n*None found*";
 ${localRows}` : "";
   const aiSection = aiOvw?.detected && aiOvw.text ? `
 ## AI Overview
-> ${truncate(aiOvw.text, 600)}` : "";
+> ${truncate(aiOvw.text, 600)}` + (aiOvw.shareUrl ? `
+
+**Shareable link:** ${aiOvw.shareUrl}` : "") : "";
   const tips = `
 ---
 \u{1F4A1} **Tips**
@@ -1106,7 +1117,7 @@ ${serpTable}${localSection}${entityIdsSection(entityIds)}${aiSection}${debugSect
         reviewCount: b.reviewCount ?? null,
         websiteUrl: b.websiteUrl ?? null
       })),
-      aiOverview: aiOvw ? { detected: aiOvw.detected === true, text: aiOvw.text ?? null } : null,
+      aiOverview: aiOvw ? { detected: aiOvw.detected === true, text: aiOvw.text ?? null, shareUrl: aiOvw.shareUrl ?? null } : null,
       entityIds: entityIds ? { entities: entityIds.entities ?? [], kgIds: entityIds.kgIds ?? [], cids: entityIds.cids ?? [], gcids: entityIds.gcids ?? [] } : null
     }
   };
@@ -1664,6 +1675,136 @@ ${rows}`,
     }
   };
 }
+function formatGoogleAdsSearch(raw, input) {
+  const parsed = parseData(raw);
+  if ("error" in parsed) return { content: [{ type: "text", text: parsed.error }], isError: true };
+  const d = parsed.data;
+  const advertisers = d.advertisers ?? [];
+  const rows = advertisers.map(
+    (a, i) => `| ${i + 1} | ${cell(a.name)} | ${cell(a.domain)} | ${a.approxAdCount ?? "\u2014"} | \`${a.advertiserId ?? "\u2014"}\` |`
+  ).join("\n");
+  const full = [
+    `# Google Ads Transparency Search: "${input.query}"`,
+    `**${advertisers.length} advertisers found**${d.region ? ` \xB7 ${d.region}` : ""}`,
+    `
+## Advertisers
+| # | Name | Domain | ~Ads | Advertiser ID |
+|---|------|--------|------|---------------|
+${rows}`,
+    `
+---
+\u{1F4A1} **Tips**
+- Pull an advertiser's ads: use \`google_ads_page_intel\` with \`advertiserId\`
+- Or pass a \`domain\` directly to \`google_ads_page_intel\``
+  ].join("\n");
+  return {
+    ...oneBlock(full),
+    structuredContent: {
+      query: input.query,
+      region: d.region ?? "US",
+      advertiserCount: advertisers.length,
+      advertisers: advertisers.map((a) => ({
+        advertiserId: a.advertiserId ?? null,
+        name: a.name ?? null,
+        domain: a.domain ?? null,
+        approxAdCount: typeof a.approxAdCount === "number" ? a.approxAdCount : null,
+        detailUrl: a.detailUrl ?? null
+      }))
+    }
+  };
+}
+function formatGoogleAdsPageIntel(raw, input) {
+  const parsed = parseData(raw);
+  if ("error" in parsed) return { content: [{ type: "text", text: parsed.error }], isError: true };
+  const d = parsed.data;
+  const ads = d.ads ?? [];
+  const rows = ads.map((a, i) => {
+    const media = a.youtubeVideoId ? `yt:${a.youtubeVideoId}` : a.videoUrl ? "video" : a.imageUrls.length ? `${a.imageUrls.length} img` : "\u2014";
+    return `| ${i + 1} | ${cell(a.format)} | ${media} | \`${a.creativeId ?? "\u2014"}\` |`;
+  }).join("\n");
+  const label = d.advertiserName ?? input.advertiserId ?? input.domain ?? "";
+  const full = [
+    `# Google Ads Transparency: ${label}`,
+    `**${ads.length} creatives** \xB7 ${d.videoCount ?? 0} video \xB7 ${d.imageCount ?? 0} image \xB7 ${d.textCount ?? 0} text${d.region ? ` \xB7 ${d.region}` : ""}`,
+    `
+## Creatives
+| # | Format | Media | Creative ID |
+|---|--------|-------|-------------|
+${rows}`,
+    `
+---
+\u{1F4A1} **Tips**
+- Transcribe a video ad: pass its \`videoUrl\` to \`google_ads_transcribe\`, or its \`youtubeVideoId\` to \`youtube_transcribe\`
+- Full ad detail is at each creative's \`detailUrl\``
+  ].join("\n");
+  return {
+    ...oneBlock(full),
+    structuredContent: {
+      advertiserId: d.advertiserId ?? null,
+      advertiserName: d.advertiserName ?? null,
+      domain: d.domain ?? null,
+      region: d.region ?? "US",
+      totalCreatives: typeof d.totalCreatives === "number" ? d.totalCreatives : ads.length,
+      videoCount: d.videoCount ?? 0,
+      imageCount: d.imageCount ?? 0,
+      textCount: d.textCount ?? 0,
+      ads: ads.map((a) => ({
+        creativeId: a.creativeId ?? null,
+        advertiserId: a.advertiserId ?? null,
+        format: a.format ?? null,
+        lastShown: a.lastShown ?? null,
+        detailUrl: a.detailUrl ?? null,
+        landingDomain: a.landingDomain ?? null,
+        imageUrls: Array.isArray(a.imageUrls) ? a.imageUrls : [],
+        youtubeVideoId: a.youtubeVideoId ?? null,
+        videoUrl: a.videoUrl ?? null,
+        variations: typeof a.variations === "number" ? a.variations : null
+      }))
+    }
+  };
+}
+function formatGoogleAdsTranscribe(raw, input) {
+  const parsed = parseData(raw);
+  if ("error" in parsed) return { content: [{ type: "text", text: parsed.error }], isError: true };
+  const d = parsed.data;
+  const text = d.text ?? "";
+  const chunks = d.chunks ?? [];
+  const durSec = d.durationMs ? (d.durationMs / 1e3).toFixed(0) : "\u2014";
+  const words = wordCount(text);
+  const chunkRows = chunks.slice(0, 50).map((c) => {
+    const sec = Number.isFinite(c.timestamp[0]) ? Math.floor(c.timestamp[0]) : 0;
+    const mm = String(Math.floor(sec / 60)).padStart(2, "0");
+    const ss = String(sec % 60).padStart(2, "0");
+    return `| ${mm}:${ss} | ${cell(truncate(c.text, 120))} |`;
+  }).join("\n");
+  const full = [
+    `# Google Ad Transcript`,
+    `**Duration:** ${durSec}s \xB7 **${words} words**`,
+    `
+## Full Transcript
+${text}`,
+    chunks.length ? `
+## Timestamped Chunks
+| Time | Text |
+|------|------|
+${chunkRows}` : "",
+    `
+---
+\u{1F4A1} Get more ads from this advertiser: use \`google_ads_page_intel\`.`
+  ].filter(Boolean).join("\n");
+  return {
+    ...oneBlock(full),
+    structuredContent: {
+      videoUrl: input.videoUrl,
+      wordCount: words,
+      chunkCount: chunks.length,
+      durationMs: typeof d.durationMs === "number" ? d.durationMs : null,
+      transcriptText: text,
+      chunks: structuredTranscriptChunks(chunks),
+      resolvedInputs: { videoUrl: input.videoUrl }
+    }
+  };
+}
 function normalizeMapsAttempts(value) {
   const attempts = Array.isArray(value) ? value : [];
   return attempts.map((attempt, index) => ({
@@ -2719,6 +2860,20 @@ var FacebookVideoTranscribeInputSchema = {
   url: z.string().url().describe("Organic Facebook reel, video, watch, post, or share URL from facebook.com, m.facebook.com, or fb.watch. The tool renders the page, extracts the best matching public Facebook CDN MP4 URL, then transcribes it. Use this when the user pastes a normal Facebook video page URL and asks for the transcript or downloadable MP4."),
   quality: z.enum(["best", "hd", "sd"]).default("best").describe("Preferred progressive MP4 quality. Use best by default; hd prefers the highest HD progressive URL; sd forces the SD URL.")
 };
+var GoogleAdsSearchInputSchema = {
+  query: z.string().min(1).describe("A domain (e.g. getviktor.com) or advertiser/brand name to look up in the Google Ads Transparency Center. Domains resolve to the advertisers running ads that point to that site."),
+  region: z.string().length(2).default("US").describe("Two-letter region code for where the ads are shown. Default US. Examples: US, CA, GB, AU."),
+  maxResults: z.number().int().min(1).max(20).default(10).describe("Maximum advertisers to return. Default 10, maximum 20.")
+};
+var GoogleAdsPageIntelInputSchema = {
+  advertiserId: z.string().optional().describe("Google Ads Transparency advertiser ID (starts with AR...). Use one returned by google_ads_search; do not construct one yourself."),
+  domain: z.string().optional().describe("A domain (e.g. getviktor.com) whose primary advertiser to inspect when advertiserId is unknown. One of advertiserId or domain is required."),
+  region: z.string().length(2).default("US").describe("Two-letter region code for where the ads are shown. Default US."),
+  maxAds: z.number().int().min(1).max(200).default(50).describe("Maximum creatives to inspect and hydrate. Default 50, maximum 200. Prefer 25-50 for focused scans.")
+};
+var GoogleAdsTranscribeInputSchema = {
+  videoUrl: z.string().url().describe("Direct googlevideo.com playback URL from a google_ads_page_intel video ad result (the videoUrl field). For YouTube-hosted ads use youtube_transcribe with the returned youtubeVideoId instead.")
+};
 var InstagramProfileContentInputSchema = {
   handle: z.string().min(1).optional().describe("Instagram handle, with or without @. Provide handle or url."),
   url: z.string().url().optional().describe("Instagram profile URL, e.g. https://www.instagram.com/nasaartemis/. Provide handle or url."),
@@ -2929,7 +3084,8 @@ var OrganicResultOutput = z.object({
 });
 var AiOverviewOutput = z.object({
   detected: z.boolean(),
-  text: NullableString
+  text: NullableString,
+  shareUrl: NullableString.optional()
 }).nullable();
 var EntityIdsOutput = z.object({
   entities: z.array(z.object({
@@ -3163,6 +3319,40 @@ var FacebookPageIntelOutputSchema = {
     variations: z.number().int().nullable()
   }))
 };
+var GoogleAdsSearchOutputSchema = {
+  query: z.string(),
+  region: z.string(),
+  advertiserCount: z.number().int().min(0),
+  advertisers: z.array(z.object({
+    advertiserId: NullableString,
+    name: NullableString,
+    domain: NullableString,
+    approxAdCount: z.number().int().nullable(),
+    detailUrl: NullableString
+  }))
+};
+var GoogleAdsPageIntelOutputSchema = {
+  advertiserId: NullableString,
+  advertiserName: NullableString,
+  domain: NullableString,
+  region: z.string(),
+  totalCreatives: z.number().int().min(0),
+  videoCount: z.number().int().min(0),
+  imageCount: z.number().int().min(0),
+  textCount: z.number().int().min(0),
+  ads: z.array(z.object({
+    creativeId: NullableString,
+    advertiserId: NullableString,
+    format: NullableString,
+    lastShown: NullableString,
+    detailUrl: NullableString,
+    landingDomain: NullableString,
+    imageUrls: z.array(z.string()),
+    youtubeVideoId: NullableString,
+    videoUrl: NullableString,
+    variations: z.number().int().nullable()
+  }))
+};
 var FacebookVideoTranscribeOutputSchema = {
   sourceUrl: z.string().url(),
   pageUrl: z.string().url(),
@@ -3305,6 +3495,17 @@ var FacebookAdTranscribeOutputSchema = {
     videoUrl: z.string().url()
   })
 };
+var GoogleAdsTranscribeOutputSchema = {
+  videoUrl: z.string().url(),
+  wordCount: z.number().int().min(0),
+  chunkCount: z.number().int().min(0),
+  durationMs: z.number().nullable(),
+  transcriptText: z.string(),
+  chunks: z.array(TranscriptChunkOutput),
+  resolvedInputs: z.object({
+    videoUrl: z.string().url()
+  })
+};
 var CaptureSerpSnapshotOutputSchema = {
   schemaVersion: z.literal("serp-intelligence.capture.v1"),
   status: z.string(),
@@ -3954,6 +4155,27 @@ function registerPaaExtractorMcpTools(server, executor, options = {}) {
     outputSchema: FacebookAdTranscribeOutputSchema,
     annotations: liveWebToolAnnotations("Facebook Ad Transcription")
   }, async (input) => formatFacebookAdTranscribe(await executor.facebookAdTranscribe(input), input));
+  server.registerTool("google_ads_search", {
+    title: "Google Ads Transparency Search",
+    description: withReportNote("Search the Google Ads Transparency Center to find advertisers running Google ads (Search, Display, YouTube) by domain or brand/advertiser name. Pass a domain like getviktor.com to see every advertiser account pointing ads at that site, or an advertiser name. Returns advertisers with their advertiser ID and an approximate ad count. Use this to discover an advertiserId, then pass it (or the domain) to google_ads_page_intel for the actual ad creatives."),
+    inputSchema: GoogleAdsSearchInputSchema,
+    outputSchema: GoogleAdsSearchOutputSchema,
+    annotations: liveWebToolAnnotations("Google Ads Transparency Search")
+  }, async (input) => formatGoogleAdsSearch(await executor.googleAdsSearch(input), input));
+  server.registerTool("google_ads_page_intel", {
+    title: "Google Ads Advertiser Intel",
+    description: withReportNote("Harvest an advertiser's ads from the Google Ads Transparency Center when the user wants a competitor's Google/YouTube ad creatives, copy angles, image URLs, or video ads. Accepts an advertiserId (from google_ads_search) or a domain directly. Returns each creative's format (text/image/video), image URLs, landing info, a detail URL, and \u2014 for video ads \u2014 the YouTube video ID and/or a direct video URL ready for transcription. Use google_ads_search first when you only have a brand name. Transcribe video ads with google_ads_transcribe (video URL) or youtube_transcribe (YouTube ID)."),
+    inputSchema: GoogleAdsPageIntelInputSchema,
+    outputSchema: GoogleAdsPageIntelOutputSchema,
+    annotations: liveWebToolAnnotations("Google Ads Advertiser Intel")
+  }, async (input) => formatGoogleAdsPageIntel(await executor.googleAdsPageIntel(input), input));
+  server.registerTool("google_ads_transcribe", {
+    title: "Google Ad Video Transcription",
+    description: "Transcribe audio from a Google video ad. Use this when google_ads_page_intel returned a direct videoUrl (a googlevideo.com playback URL) for a video creative and the user asks what the ad says or wants its claims/messaging. Returns full transcript, timestamped chunks, and word count. For YouTube-hosted ads, pass the returned youtubeVideoId to youtube_transcribe instead; run google_ads_page_intel first when you only have an advertiser/domain.",
+    inputSchema: GoogleAdsTranscribeInputSchema,
+    outputSchema: GoogleAdsTranscribeOutputSchema,
+    annotations: liveWebToolAnnotations("Google Ad Video Transcription")
+  }, async (input) => formatGoogleAdsTranscribe(await executor.googleAdsTranscribe(input), input));
   server.registerTool("facebook_video_transcribe", {
     title: "Facebook Organic Video Transcription",
     description: withReportNote("Transcribe audio from an organic Facebook reel, video, watch, post, or share URL, including fb.watch links. Use this when the user pastes a normal Facebook video page URL and wants the transcript or downloadable MP4. Renders the Facebook page in a browser, selects the best matching public Facebook CDN MP4 URL from page state, then returns sourceUrl, resolved pageUrl, videoId, ownerName, selectedQuality, bitrate, videoDurationSec, extracted MP4 URL, full transcript, and timestamped chunks."),
@@ -4273,6 +4495,15 @@ var HttpMcpToolExecutor = class {
   facebookVideoTranscribe(input) {
     return this.call("/facebook/video-transcribe", input, this.httpTimeoutOverrideMs ?? 24e4);
   }
+  googleAdsSearch(input) {
+    return this.call("/google-ads/search", input, this.httpTimeoutOverrideMs ?? 24e4);
+  }
+  googleAdsPageIntel(input) {
+    return this.call("/google-ads/page-intel", input, this.httpTimeoutOverrideMs ?? 3e5);
+  }
+  googleAdsTranscribe(input) {
+    return this.call("/google-ads/transcribe", input, this.httpTimeoutOverrideMs ?? 24e4);
+  }
   instagramProfileContent(input) {
     return this.call("/instagram/profile-content", withDefaultHostedProfile(input), this.httpTimeoutOverrideMs ?? 24e4);
   }
@@ -5817,4 +6048,4 @@ export {
   exportFanout,
   registerBrowserAgentMcpTools
 };
-//# sourceMappingURL=chunk-NGD4ITAO.js.map
+//# sourceMappingURL=chunk-KPF64WST.js.map