mcp-scraper 0.2.14 → 0.2.15

package/dist/bin/mcp-scraper-cli.js

@@ -9,7 +9,7 @@ import {
 } from "../chunk-F44RBOJ5.js";
 import {
   PACKAGE_VERSION
-} from "../chunk-ROCXJOVY.js";
+} from "../chunk-2CQXHSWC.js";
 
 // src/cli/human-cli.ts
 import { Command } from "commander";

package/dist/bin/mcp-scraper-combined-stdio-server.cjs

@@ -108,6 +108,9 @@ var HttpMcpToolExecutor = class {
   facebookAdTranscribe(input) {
     return this.call("/facebook/transcribe", input);
   }
+  facebookVideoTranscribe(input) {
+    return this.call("/facebook/video-transcribe", input, this.httpTimeoutOverrideMs ?? 24e4);
+  }
   mapsPlaceIntel(input) {
     return this.call("/maps/place", input);
   }
@@ -138,7 +141,7 @@ var import_node_os2 = require("os");
 var import_node_path2 = require("path");
 
 // src/version.ts
-var PACKAGE_VERSION = "0.2.14";
+var PACKAGE_VERSION = "0.2.15";
 
 // src/mcp/browser-agent-tool-schemas.ts
 var import_zod = require("zod");
@@ -1918,6 +1921,62 @@ ${chunkRows}` : "",
   ].filter(Boolean).join("\n");
   return oneBlock(full);
 }
+function formatFacebookVideoTranscribe(raw, input) {
+  const parsed = parseData(raw);
+  if ("error" in parsed) return { content: [{ type: "text", text: parsed.error }], isError: true };
+  const d = parsed.data;
+  const text = d.text ?? "";
+  const chunks = d.chunks ?? [];
+  const wordCount = text.trim() ? text.trim().split(/\s+/).length : 0;
+  const durSec = d.durationMs ? (d.durationMs / 1e3).toFixed(0) : "\u2014";
+  const videoDuration = typeof d.videoDurationSec === "number" ? `${Math.round(d.videoDurationSec)}s` : "\u2014";
+  const label = d.videoId ? `Facebook Organic Video \`${d.videoId}\`` : "Facebook Organic Video";
+  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 = [
+    `# ${label} Transcript`,
+    d.ownerName ? `**Owner:** ${d.ownerName}` : "",
+    `**Video duration:** ${videoDuration} \xB7 **Transcribed in:** ${durSec}s \xB7 **${wordCount} words**`,
+    d.pageUrl ? `**Page URL:** ${d.pageUrl}` : `**Page URL:** ${input.url}`,
+    d.videoUrl ? `**Extracted MP4:** \`${d.videoUrl}\`` : "",
+    `
+## Full Transcript
+${text}`,
+    chunks.length ? `
+## Timestamped Chunks
+| Time | Text |
+|------|------|
+${chunkRows}` : "",
+    `
+---
+\u{1F4A1} To download the extracted MP4, call the Facebook media endpoint with the extracted MP4 URL.`
+  ].filter(Boolean).join("\n");
+  return {
+    ...oneBlock(full),
+    structuredContent: {
+      sourceUrl: d.sourceUrl ?? input.url,
+      pageUrl: d.pageUrl ?? input.url,
+      videoId: d.videoId ?? null,
+      ownerName: d.ownerName ?? null,
+      selectedQuality: d.selectedQuality ?? input.quality ?? "best",
+      bitrate: typeof d.bitrate === "number" ? d.bitrate : null,
+      videoDurationSec: typeof d.videoDurationSec === "number" ? d.videoDurationSec : null,
+      videoUrl: d.videoUrl ?? "",
+      wordCount,
+      chunkCount: chunks.length,
+      transcriptText: text,
+      chunks: chunks.map((c) => ({
+        startSec: Number.isFinite(c.timestamp[0]) ? c.timestamp[0] : 0,
+        endSec: Number.isFinite(c.timestamp[1]) ? c.timestamp[1] : 0,
+        text: c.text
+      }))
+    }
+  };
+}
 
 // src/mcp/mcp-tool-schemas.ts
 var import_zod2 = require("zod");
@@ -1973,6 +2032,10 @@ var FacebookAdSearchInputSchema = {
 var FacebookAdTranscribeInputSchema = {
   videoUrl: import_zod2.z.string().url().describe("Facebook CDN video URL from a facebook_page_intel result")
 };
+var FacebookVideoTranscribeInputSchema = {
+  url: import_zod2.z.string().url().describe("Organic Facebook reel, video, watch, or share URL. The tool renders the page, extracts the best public Facebook CDN MP4 URL, then transcribes it."),
+  quality: import_zod2.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 MapsPlaceIntelInputSchema = {
   businessName: import_zod2.z.string().min(1).describe('Business name only. If user says "Elite Roofing Denver CO", use businessName="Elite Roofing" and location="Denver, CO".'),
   location: import_zod2.z.string().min(1).describe('City/region/country where the business should be searched, e.g. "Denver, CO". Infer from the user request when possible.'),
@@ -2338,6 +2401,24 @@ var FacebookPageIntelOutputSchema = {
     variations: import_zod2.z.number().int().nullable()
   }))
 };
+var FacebookVideoTranscribeOutputSchema = {
+  sourceUrl: import_zod2.z.string().url(),
+  pageUrl: import_zod2.z.string().url(),
+  videoId: NullableString,
+  ownerName: NullableString,
+  selectedQuality: import_zod2.z.string(),
+  bitrate: import_zod2.z.number().int().nullable(),
+  videoDurationSec: import_zod2.z.number().nullable(),
+  videoUrl: import_zod2.z.string().url(),
+  wordCount: import_zod2.z.number().int().min(0),
+  chunkCount: import_zod2.z.number().int().min(0),
+  transcriptText: import_zod2.z.string(),
+  chunks: import_zod2.z.array(import_zod2.z.object({
+    startSec: import_zod2.z.number(),
+    endSec: import_zod2.z.number(),
+    text: import_zod2.z.string()
+  }))
+};
 var CreditsInfoInputSchema = {
   item: import_zod2.z.string().optional().describe('Optional tool, action, or feature to look up, e.g. "maps reviews", "extract_url", "YouTube transcription", or "concurrency"'),
   includeLedger: import_zod2.z.boolean().default(false).describe("Whether to include recent credit ledger entries")
@@ -2832,6 +2913,13 @@ function registerPaaExtractorMcpTools(server2, executor, options = {}) {
     inputSchema: FacebookAdTranscribeInputSchema,
     annotations: liveWebToolAnnotations("Facebook Ad Transcription")
   }, async (input) => formatFacebookAdTranscribe(await executor.facebookAdTranscribe(input), input));
+  server2.registerTool("facebook_video_transcribe", {
+    title: "Facebook Organic Video Transcription",
+    description: withReportNote("Transcribe audio from an organic Facebook reel, video, watch, or share URL. Renders the Facebook page, extracts the best public Facebook CDN MP4 URL from the page state, then returns the full transcript, timestamped chunks, and extracted MP4 URL."),
+    inputSchema: FacebookVideoTranscribeInputSchema,
+    outputSchema: FacebookVideoTranscribeOutputSchema,
+    annotations: liveWebToolAnnotations("Facebook Organic Video Transcription")
+  }, async (input) => formatFacebookVideoTranscribe(await executor.facebookVideoTranscribe(input), input));
   server2.registerTool("maps_place_intel", {
     title: "Google Maps Business Profile Details",
     description: withReportNote('Extract Google Maps business intelligence for one known/named business: rating, review count, category, address, phone, website, hours, booking URL, review histogram, review topics, about attributes, entity IDs, and optional review cards. Do not use this for category searches, local market prospect lists, or requests for multiple GMB/GBP profiles; use maps_search first for those. Split business name from location (e.g. "Elite Roofing Denver CO" => businessName "Elite Roofing", location "Denver, CO"). Pass includeReviews true when the user asks for reviews/customer pain.'),