mcp-scraper 0.26.3 → 0.28.0

package/dist/{server-GN34MGAE.js → server-QQB4BNAB.js}

@@ -12,9 +12,11 @@ import {
   MapsSelectors,
   buildYouTubeChannelVideosUrl,
   deleteKernelProxyId,
+  encodeUule,
   harvest,
+  normalizeLocation,
   resolveKernelProxyId
-} from "./chunk-E5J4HJBO.js";
+} from "./chunk-YH5DRFWQ.js";
 import {
   HttpMcpToolExecutor,
   assessTranscriptSignal,
@@ -34,7 +36,7 @@ import {
   sanitizeAttempts,
   sanitizeHarvestResult,
   transcribeMediaUrl
-} from "./chunk-73YM3DEB.js";
+} from "./chunk-LTT7QX3C.js";
 import {
   auditImages,
   buildLinkReport,
@@ -63,7 +65,7 @@ import {
   workflowOutputBaseDir,
   workflowStepCount,
   workflowSupportsSteps
-} from "./chunk-FUWZWKGO.js";
+} from "./chunk-KAMKTSJQ.js";
 import "./chunk-K443GQY5.js";
 import {
   DEFAULT_MAPS_PROXY_MODE,
@@ -74,8 +76,8 @@ import {
   RawMapsHoursRowSchema,
   RawMapsOverviewSchema,
   RawMapsReviewStatsSchema
-} from "./chunk-XGIPATLV.js";
-import "./chunk-27VXTOD2.js";
+} from "./chunk-ZQSHKWV4.js";
+import "./chunk-QG77M3YJ.js";
 import {
   completeExtractJob,
   countSuccessfulPages,
@@ -5061,7 +5063,7 @@ async function extractSite(opts) {
 }
 
 // src/api/server.ts
-import { Hono as Hono24 } from "hono";
+import { Hono as Hono25 } from "hono";
 import { serve as serveInngest } from "inngest/hono";
 
 // src/inngest/client.ts
@@ -13077,6 +13079,262 @@ videoApp.post("/status", createApiKeyAuth(), async (c) => {
 // src/api/maps-routes.ts
 import { Hono as Hono11 } from "hono";
 
+// src/extractor/MapsSearchExtractor.ts
+var LOCAL_RESULTS_PAGE_LIMIT = 8;
+var LOCAL_RESULTS_WAIT_MS = 1e3;
+function buildGoogleLocalResultsUrl(input) {
+  const query = [input.query, input.location].filter(Boolean).join(" ");
+  const params = new URLSearchParams({ q: query, udm: "1", gl: input.gl, hl: input.hl, pws: "0" });
+  if (input.location) params.set("uule", encodeUule(normalizeLocation(input.location)));
+  return `https://www.google.com/search?${params.toString()}`;
+}
+var MapsSearchExtractor = class {
+  constructor(driver) {
+    this.driver = driver;
+  }
+  driver;
+  async extract(options) {
+    const startMs = Date.now();
+    const searchQuery = [options.query, options.location].filter(Boolean).join(" ");
+    const searchUrl = buildGoogleLocalResultsUrl(options);
+    const config = {
+      headless: options.headless,
+      kernelApiKey: options.kernelApiKey,
+      kernelProxyId: options.kernelProxyId,
+      kernelProxyResolution: options.kernelProxyResolution,
+      proxyMode: options.proxyMode,
+      viewport: { width: 1280, height: 900 },
+      locale: `${options.hl}-${options.gl.toUpperCase()}`,
+      debug: options.debug
+    };
+    try {
+      await this.driver.launch(config);
+      const page = this.driver.getPage();
+      await page.goto(searchUrl, { waitUntil: "domcontentloaded", timeout: 6e4 });
+      await page.waitForTimeout(LOCAL_RESULTS_WAIT_MS);
+      if (await this.detectBlock(page)) throw new CaptchaError(RECAPTCHA_INSTRUCTIONS);
+      const results = await this.collectResults(page, options);
+      return {
+        query: options.query,
+        location: options.location ?? null,
+        searchQuery,
+        searchUrl,
+        extractedAt: (/* @__PURE__ */ new Date()).toISOString(),
+        requestedMaxResults: options.maxResults,
+        resultCount: results.length,
+        results,
+        durationMs: Date.now() - startMs
+      };
+    } finally {
+      await this.driver.close();
+    }
+  }
+  async detectBlock(page) {
+    return page.evaluate(() => {
+      const text2 = document.body.innerText.slice(0, 2e3);
+      return /unusual traffic|captcha|recaptcha|about this page/i.test(text2) || /\/sorry\//.test(location.href);
+    });
+  }
+  async collectResults(page, options) {
+    const seen = /* @__PURE__ */ new Set();
+    const results = [];
+    let pageCount = 0;
+    while (results.length < options.maxResults && pageCount < LOCAL_RESULTS_PAGE_LIMIT) {
+      if (await this.detectBlock(page)) throw new CaptchaError(RECAPTCHA_INSTRUCTIONS);
+      const cards = await this.extractVisibleCards(page, options.location);
+      for (const card of cards) {
+        if (results.length >= options.maxResults || seen.has(card.cardKey)) continue;
+        seen.add(card.cardKey);
+        const details = await this.openCardAndExtractDialog(page, card, options.includeServices);
+        results.push({
+          position: results.length + 1,
+          name: card.name,
+          placeUrl: details?.placeUrl ?? card.placeUrl,
+          cid: details?.cid ?? card.cid,
+          cidDecimal: details?.cidDecimal ?? card.cidDecimal,
+          rating: details?.rating ?? card.rating,
+          reviewCount: details?.reviewCount ?? card.reviewCount,
+          category: details?.category ?? card.category,
+          address: details?.address ?? card.address,
+          phone: details?.phone ?? card.phone,
+          hoursStatus: details?.hoursStatus ?? card.hoursStatus,
+          websiteUrl: details?.websiteUrl ?? card.websiteUrl,
+          directionsUrl: card.directionsUrl,
+          metadata: card.metadata,
+          services: details?.services ?? [],
+          areasServed: details?.areasServed ?? [],
+          profileDetailsStatus: details ? "collected" : "unavailable"
+        });
+      }
+      if (results.length >= options.maxResults) break;
+      const advanced = await this.advanceToMoreResults(page);
+      if (!advanced) break;
+      pageCount += 1;
+    }
+    return results;
+  }
+  async extractVisibleCards(page, location2) {
+    const fallbackLocation = location2 ?? "";
+    return page.evaluate((locationHint) => {
+      const normalize = (value) => {
+        const text2 = value?.replace(/\s+/g, " ").trim() ?? "";
+        return text2 || null;
+      };
+      const phonePattern = /(?:\+?1[\s.-]?)?\(?\d{3}\)?[\s.-]\d{3}[\s.-]\d{4}/;
+      const addressPattern = /\b[A-Z]{2}\s+\d{5}(?:-\d{4})?\b|\b(?:St|Street|Ave|Avenue|Rd|Road|Blvd|Boulevard|Dr|Drive|Ln|Lane|Way|Pkwy|Parkway)\b/i;
+      const cards = Array.from(document.querySelectorAll("button")).filter((button) => {
+        const name = normalize(button.querySelector("h3")?.textContent);
+        return Boolean(name) && !button.closest('[role="dialog"], dialog') && !/\bsponsored\b/i.test(button.innerText);
+      });
+      const seen = /* @__PURE__ */ new Set();
+      const out = [];
+      for (const [cardIndex, card] of cards.entries()) {
+        const name = normalize(card.querySelector("h3")?.textContent);
+        if (!name) continue;
+        const text2 = normalize(card.innerText) ?? name;
+        const key = `${name.toLowerCase()}|${text2.toLowerCase()}`;
+        if (seen.has(key)) continue;
+        seen.add(key);
+        const lines = text2.split(/\n+/).map((line) => normalize(line)).filter((line) => Boolean(line));
+        const aria = normalize(card.getAttribute("aria-label"));
+        const rating = (aria ?? text2).match(/(?:rated\s+)?(\d(?:\.\d)?)\s*(?:out of 5)?/i)?.[1] ?? null;
+        const reviewCount = (aria ?? text2).match(/(?:\(|·\s*)([\d,]+)\s+(?:user\s+)?reviews?\b/i)?.[1]?.replace(/,/g, "") ?? null;
+        const phone = lines.find((line) => phonePattern.test(line)) ?? null;
+        const address = lines.find((line) => addressPattern.test(line)) ?? null;
+        const hoursStatus = lines.find((line) => /\b(?:open|closed|opens|closes)\b/i.test(line)) ?? null;
+        const category = lines.find(
+          (line) => line !== name && line !== phone && line !== address && line !== hoursStatus && !/^(?:rated|\d[\d.,]*\s*(?:out of 5|reviews?)?)\b/i.test(line) && !/sponsored|website|directions|call|save|share|more/i.test(line)
+        ) ?? null;
+        const mapsSearchUrl = `https://www.google.com/maps/search/${encodeURIComponent([name, locationHint].filter(Boolean).join(" "))}`;
+        out.push({
+          cardIndex,
+          cardKey: key,
+          name,
+          placeUrl: mapsSearchUrl,
+          cid: null,
+          cidDecimal: null,
+          rating,
+          reviewCount,
+          category,
+          address,
+          phone,
+          hoursStatus,
+          websiteUrl: null,
+          directionsUrl: `https://www.google.com/maps/dir/?api=1&destination=${encodeURIComponent([name, address].filter(Boolean).join(", "))}`,
+          metadata: lines.slice(0, 20)
+        });
+      }
+      return out;
+    }, fallbackLocation);
+  }
+  async openCardAndExtractDialog(page, card, includeServices) {
+    const clicked = await page.evaluate((input) => {
+      const normalize = (value) => (value ?? "").replace(/\s+/g, " ").trim().toLowerCase();
+      const cards = Array.from(document.querySelectorAll("button")).filter(
+        (button) => Boolean(button.querySelector("h3")) && !button.closest('[role="dialog"], dialog') && !/\bsponsored\b/i.test(button.innerText)
+      );
+      const exact = cards.find((button) => `${normalize(button.querySelector("h3")?.textContent)}|${normalize(button.innerText)}` === input.cardKey);
+      const target = exact ?? cards[input.cardIndex];
+      if (!target) return false;
+      target.click();
+      return true;
+    }, { cardIndex: card.cardIndex, cardKey: card.cardKey });
+    if (!clicked) return null;
+    const dialog = page.locator('dialog, [role="dialog"]').filter({ has: page.locator('h1, h2, [role="heading"]') }).last();
+    await dialog.waitFor({ state: "visible", timeout: 6e3 }).catch(() => {
+    });
+    if (await dialog.count() === 0 || !await dialog.isVisible().catch(() => false)) return null;
+    const details = await dialog.evaluate((root, includeServicesValue) => {
+      const text2 = (value) => value?.replace(/\s+/g, " ").trim() ?? "";
+      const phonePattern = /(?:\+?1[\s.-]?)?\(?\d{3}\)?[\s.-]\d{3}[\s.-]\d{4}/;
+      const buttons = Array.from(root.querySelectorAll("button"));
+      const buttonTexts = buttons.map((button) => text2(button.innerText)).filter(Boolean);
+      const servicesText = includeServicesValue ? buttonTexts.find((value) => /^services\s*:/i.test(value)) : null;
+      const services = servicesText ? [...new Set(servicesText.replace(/^services\s*:\s*/i, "").split(/\s*[,·]\s*/).map((value) => value.trim()).filter(Boolean))] : [];
+      const areaText = includeServicesValue ? buttonTexts.find((value) => /^serves\b/i.test(value)) : null;
+      const areasServed = areaText ? [areaText.replace(/^serves\s*/i, "").trim()].filter(Boolean) : [];
+      const externalLinks = Array.from(root.querySelectorAll("a[href]"));
+      const website = externalLinks.find((link) => /^website$/i.test(text2(link.innerText)) || /^website$/i.test(link.getAttribute("aria-label") ?? ""));
+      const mapLink = externalLinks.find((link) => /google\.[^/]+\/maps|\/maps\//i.test(link.href));
+      const addressLink = externalLinks.find((link) => /\b[A-Z]{2}\s+\d{5}(?:-\d{4})?\b|\b(?:St|Street|Ave|Avenue|Rd|Road|Blvd|Boulevard|Dr|Drive|Ln|Lane|Way|Pkwy|Parkway)\b/i.test(text2(link.innerText)));
+      const allText = text2(root.innerText);
+      const rating = allText.match(/\b(?:rated\s+)?(\d(?:\.\d)?)\s+out of 5\b/i)?.[1] ?? allText.match(/\b(\d(?:\.\d)?)\s+rated\b/i)?.[1] ?? null;
+      const reviewCount = allText.match(/([\d,]+)\s+(?:user\s+)?reviews?\b/i)?.[1]?.replace(/,/g, "") ?? null;
+      const phoneButton = buttonTexts.find((value) => /^call\b/i.test(value) || phonePattern.test(value)) ?? "";
+      const phone = phoneButton.match(phonePattern)?.[0] ?? null;
+      const hoursStatus = buttonTexts.find((value) => /\b(?:open|closed|opens|closes)\b/i.test(value)) ?? null;
+      const placeUrl = mapLink?.href ?? addressLink?.href ?? location.href;
+      const fid = placeUrl.match(/!1s(0x[0-9a-f]+):(0x[0-9a-f]+)/i);
+      const cid = fid ? `${fid[1]}:${fid[2]}` : null;
+      let cidDecimal = null;
+      if (fid) {
+        try {
+          cidDecimal = BigInt(fid[2]).toString();
+        } catch {
+        }
+      }
+      if (!cidDecimal) {
+        const cidFromQuery = new URL(placeUrl).searchParams.get("cid");
+        cidDecimal = cidFromQuery && /^\d+$/.test(cidFromQuery) ? cidFromQuery : null;
+      }
+      return {
+        placeUrl,
+        cid,
+        cidDecimal,
+        rating,
+        reviewCount,
+        category: null,
+        address: text2(addressLink?.innerText) || null,
+        phone,
+        websiteUrl: website?.href ?? null,
+        hoursStatus,
+        services,
+        areasServed
+      };
+    }, includeServices);
+    await this.closeDialog(page);
+    return details;
+  }
+  async closeDialog(page) {
+    const closed = await page.evaluate(() => {
+      const dialogs = Array.from(document.querySelectorAll('dialog, [role="dialog"]'));
+      const dialog = dialogs.find((candidate) => candidate.offsetParent !== null);
+      const close = dialog && Array.from(dialog.querySelectorAll("button")).find(
+        (button) => /^close$/i.test(button.getAttribute("aria-label") ?? "") || /^close$/i.test(button.innerText.trim())
+      );
+      close?.click();
+      return Boolean(close);
+    });
+    if (closed) await page.waitForTimeout(250);
+  }
+  async advanceToMoreResults(page) {
+    const before = await page.evaluate(() => document.body.innerText.slice(-2e3));
+    const advanced = await page.evaluate(() => {
+      const visible = (element) => element.offsetParent !== null;
+      const mobile = Array.from(document.querySelectorAll("button, a")).find(
+        (element) => visible(element) && /^(more search results|more results)$/i.test(element.innerText.trim())
+      );
+      if (mobile) {
+        mobile.click();
+        return true;
+      }
+      const next = Array.from(document.querySelectorAll("a[href]")).find(
+        (link) => visible(link) && (/^next$/i.test(link.innerText.trim()) || /[?&]start=\d+/.test(link.href))
+      );
+      if (next) {
+        next.click();
+        return true;
+      }
+      return false;
+    });
+    if (!advanced) return false;
+    await page.waitForTimeout(LOCAL_RESULTS_WAIT_MS);
+    await page.waitForFunction((previous) => document.body.innerText.slice(-2e3) !== previous, before, { timeout: 5e3 }).catch(() => {
+    });
+    return true;
+  }
+};
+
 // src/extractor/MapsNavigator.ts
 var MapsNavigator = class {
   constructor(page) {
@@ -13135,32 +13393,26 @@ var MapsNavigator = class {
     await this.page.waitForSelector('[role="main"] h2', { timeout: 5e3 }).catch(() => {
     });
   }
-  async navigateToSerpBusinessPanel(businessName, location2) {
-    const query = `${businessName} ${location2}`;
-    const searchUrl = `https://www.google.com/search?q=${encodeURIComponent(query)}&udm=1`;
+  async navigateToSerpBusinessPanel(businessName, location2, gl, hl) {
+    const searchUrl = buildGoogleLocalResultsUrl({ query: businessName, location: location2, gl, hl });
     await this.page.goto(searchUrl, { waitUntil: "domcontentloaded", timeout: 45e3 }).catch(() => {
     });
-    await this.page.waitForTimeout(2e3);
+    await this.page.waitForTimeout(1e3);
     const clicked = await this.page.evaluate((name) => {
-      const trimmedName = name.trim();
-      const words = trimmedName.split(/\s+/);
-      const prefix = words.slice(0, Math.min(2, words.length)).join(" ");
-      const headings = Array.from(document.querySelectorAll('h3, div[role="heading"]'));
-      const heading = headings.find((el2) => el2.innerText?.trim() === trimmedName) ?? headings.find((el2) => el2.innerText?.trim().startsWith(prefix));
-      const clickable = heading?.closest('a, div[role="link"], div[jsaction]');
-      if (clickable) {
-        clickable.click();
-        return true;
-      }
-      if (heading) {
-        heading.click();
+      const normalized = name.replace(/\s+/g, " ").trim().toLowerCase();
+      const card = Array.from(document.querySelectorAll("button")).find(
+        (button) => button.querySelector("h3")?.textContent?.replace(/\s+/g, " ").trim().toLowerCase() === normalized
+      );
+      if (card) {
+        card.click();
         return true;
       }
       return false;
     }, businessName);
     if (!clicked) return false;
-    await this.page.waitForTimeout(2e3);
-    return true;
+    await this.page.waitForSelector('dialog, [role="dialog"]', { timeout: 5e3 }).catch(() => {
+    });
+    return this.page.locator('dialog, [role="dialog"]').count().then((count) => count > 0);
   }
   buildReviewsUrl(placeUrl) {
     const fid = placeUrl.match(/!1s(0x[0-9a-f]+:0x[0-9a-f]+)/i);
@@ -13336,7 +13588,7 @@ var MapsExtractor = class {
       let servicesStatus = "not_requested";
       if (options.includeServices) {
         try {
-          const opened = await navigator.navigateToSerpBusinessPanel(options.businessName, options.location);
+          const opened = await navigator.navigateToSerpBusinessPanel(options.businessName, options.location, options.gl, options.hl);
           if (opened) {
             const result = await this.extractServicesAndAreasServed(page);
             services = result.services;
@@ -13557,242 +13809,15 @@ var MapsExtractor = class {
     return result.data;
   }
   async extractServicesAndAreasServed(page) {
-    await page.waitForTimeout(1500);
-    let services = [];
-    if (await this.clickRowStartingWithRetry(page, "Services:")) {
-      await page.waitForSelector('[role="dialog"]', { timeout: 5e3 }).catch(() => {
-      });
-      await page.waitForTimeout(800);
-      services = await this.extractOpenDialogItems(page, "Services");
-      await this.closeOpenDialog(page, "Services");
-      await page.waitForTimeout(500);
-    }
-    let areasServed = [];
-    if (await this.clickRowStartingWithRetry(page, "Serves ")) {
-      await page.waitForSelector('[role="dialog"]', { timeout: 5e3 }).catch(() => {
-      });
-      await page.waitForTimeout(800);
-      areasServed = await this.extractOpenDialogItems(page, "Areas Served");
-      await this.closeOpenDialog(page, "Areas Served");
-    }
-    return { services, areasServed };
-  }
-  async clickRowStartingWithRetry(page, prefix) {
-    if (await this.clickRowStartingWith(page, prefix)) return true;
-    await page.waitForTimeout(2e3);
-    return this.clickRowStartingWith(page, prefix);
-  }
-  async clickRowStartingWith(page, prefix) {
-    return page.evaluate((p) => {
-      const el2 = Array.from(document.querySelectorAll('button, a, div[role="button"]')).find(
-        (e) => (e.innerText ?? "").trim().startsWith(p)
-      );
-      if (!el2) return false;
-      el2.click();
-      return true;
-    }, prefix);
-  }
-  async extractOpenDialogItems(page, expectedHeading) {
-    return page.evaluate((expected) => {
-      const dialogs = Array.from(document.querySelectorAll('[role="dialog"]'));
-      const withHeading = dialogs.find((d) => {
-        const h = d.querySelector('h1, h2, [role="heading"]');
-        const headingText = h?.innerText?.trim() ?? "";
-        return headingText.startsWith(expected);
-      });
-      if (!withHeading) return [];
-      const headingEl = withHeading.querySelector('h1, h2, [role="heading"]');
-      const headingLines = new Set(
-        (headingEl?.innerText ?? "").split("\n").map((s) => s.trim()).filter(Boolean)
-      );
-      const seen = /* @__PURE__ */ new Set();
-      const items = [];
-      withHeading.querySelectorAll("*").forEach((el2) => {
-        if (el2.children.length > 0) return;
-        const text2 = el2.innerText?.trim() ?? el2.textContent?.trim() ?? "";
-        if (!text2 || headingLines.has(text2)) return;
-        if (/^(back|close|done|thanks)$/i.test(text2)) return;
-        if (/^serves\s.*following/i.test(text2)) return;
-        if (seen.has(text2)) return;
-        seen.add(text2);
-        items.push(text2);
-      });
-      return items;
-    }, expectedHeading);
-  }
-  async closeOpenDialog(page, expectedHeading) {
-    await page.evaluate((expected) => {
-      const dialog = Array.from(document.querySelectorAll('[role="dialog"]')).find((d) => {
-        const h = d.querySelector('h1, h2, [role="heading"]');
-        const headingText = h?.innerText?.trim() ?? "";
-        return headingText.startsWith(expected);
-      });
-      const closeBtn = dialog?.querySelector('[aria-label*="Close" i], button[aria-label*="close" i]');
-      closeBtn?.click();
-    }, expectedHeading);
-  }
-};
-
-// src/extractor/MapsSearchExtractor.ts
-var MAPS_SEARCH_SCROLL_BUDGET_MS = 6e4;
-var MAPS_SEARCH_SCROLL_STEP_MS = 1200;
-var MAPS_SEARCH_MAX_NO_GROWTH_ROUNDS = 4;
-var MapsSearchExtractor = class {
-  constructor(driver) {
-    this.driver = driver;
-  }
-  driver;
-  async extract(options) {
-    const startMs = Date.now();
-    const searchQuery = [options.query, options.location].filter(Boolean).join(" ");
-    const searchUrl = `https://www.google.com/maps/search/${encodeURIComponent(searchQuery)}?hl=${encodeURIComponent(options.hl)}`;
-    const config = {
-      headless: options.headless,
-      kernelApiKey: options.kernelApiKey,
-      kernelProxyId: options.kernelProxyId,
-      kernelProxyResolution: options.kernelProxyResolution,
-      proxyMode: options.proxyMode,
-      viewport: { width: 1280, height: 900 },
-      locale: `${options.hl}-${options.gl.toUpperCase()}`,
-      debug: options.debug
-    };
-    try {
-      await this.driver.launch(config);
-      const page = this.driver.getPage();
-      await page.goto(searchUrl, { waitUntil: "domcontentloaded", timeout: 6e4 });
-      await page.waitForTimeout(3e3);
-      const blocked = await this.detectBlock(page);
-      if (blocked) throw new CaptchaError(RECAPTCHA_INSTRUCTIONS);
-      const results = await this.collectResults(page, options.maxResults);
-      return {
-        query: options.query,
-        location: options.location ?? null,
-        searchQuery,
-        searchUrl,
-        extractedAt: (/* @__PURE__ */ new Date()).toISOString(),
-        requestedMaxResults: options.maxResults,
-        resultCount: results.length,
-        results,
-        durationMs: Date.now() - startMs
-      };
-    } finally {
-      await this.driver.close();
-    }
-  }
-  async detectBlock(page) {
     return page.evaluate(() => {
-      const text2 = document.body.innerText.slice(0, 2e3);
-      return /unusual traffic|captcha|recaptcha|about this page/i.test(text2) || /\/sorry\//.test(location.href);
-    });
-  }
-  async collectResults(page, maxResults) {
-    const seen = /* @__PURE__ */ new Map();
-    const started = Date.now();
-    let noGrowthRounds = 0;
-    while (Date.now() - started < MAPS_SEARCH_SCROLL_BUDGET_MS) {
-      const before = seen.size;
-      const batch = await this.extractVisibleResults(page);
-      for (const result of batch) {
-        const key = this.resultKey(result);
-        if (!seen.has(key)) seen.set(key, { ...result, position: seen.size + 1 });
-        if (seen.size >= maxResults) break;
-      }
-      if (seen.size >= maxResults) break;
-      if (seen.size === before) noGrowthRounds += 1;
-      else noGrowthRounds = 0;
-      if (noGrowthRounds >= MAPS_SEARCH_MAX_NO_GROWTH_ROUNDS) break;
-      await page.evaluate(() => {
-        const feed = document.querySelector('[role="feed"]');
-        if (feed) {
-          feed.scrollTop = feed.scrollHeight;
-        } else {
-          window.scrollTo(0, document.body.scrollHeight);
-        }
-      });
-      await page.waitForTimeout(MAPS_SEARCH_SCROLL_STEP_MS);
-    }
-    return [...seen.values()].slice(0, maxResults);
-  }
-  resultKey(result) {
-    return result.cidDecimal ?? result.placeUrl.replace(/[?&].*$/, "") ?? result.name;
-  }
-  async extractVisibleResults(page) {
-    return page.evaluate(() => {
-      function normalizeText(value) {
-        const text2 = value?.replace(/\s+/g, " ").trim() ?? "";
-        return text2 || null;
-      }
-      function cidFromUrl(url) {
-        const fid = url.match(/!1s(0x[0-9a-f]+):(0x[0-9a-f]+)/i);
-        if (!fid) return { cid: null, cidDecimal: null };
-        let cidDecimal = null;
-        try {
-          cidDecimal = BigInt(fid[2]).toString();
-        } catch {
-        }
-        return { cid: `${fid[1]}:${fid[2]}`, cidDecimal };
-      }
-      function textParts(card) {
-        if (!card) return [];
-        const parts = [];
-        card.querySelectorAll("div, span").forEach((el2) => {
-          const text2 = Array.from(el2.childNodes).filter((node) => node.nodeType === 3).map((node) => node.textContent?.trim() ?? "").filter((text3) => text3.length > 1 && text3.length < 140).join(" ");
-          if (text2 && !parts.includes(text2)) parts.push(text2);
-        });
-        return parts;
-      }
-      function firstMatching(parts, pattern) {
-        const value = parts.find((part) => pattern.test(part));
-        return value ?? null;
-      }
-      function normalizedSet(values) {
-        return new Set(values.filter(Boolean).map((value) => value.toLowerCase()));
-      }
-      const out = [];
-      const seen = /* @__PURE__ */ new Set();
-      const anchors = Array.from(document.querySelectorAll('a[href*="/maps/place/"]'));
-      for (const anchor of anchors) {
-        const placeUrl = anchor.href;
-        const stableUrl = placeUrl.replace(/[?&].*$/, "");
-        if (seen.has(stableUrl)) continue;
-        seen.add(stableUrl);
-        const card = anchor.closest('.Nv2PK, [role="article"], .bfdHYd') ?? anchor.parentElement;
-        const parts = textParts(card);
-        const aria = normalizeText(anchor.getAttribute("aria-label"));
-        const heading = normalizeText(card?.querySelector('.qBF1Pd, .fontHeadlineSmall, [role="heading"]')?.textContent);
-        const name = aria ?? heading ?? parts[0] ?? stableUrl;
-        const links = Array.from(card?.querySelectorAll("a[href]") ?? []);
-        const websiteUrl = links.find((link) => link.href.startsWith("http") && !link.href.includes("google."))?.href ?? null;
-        const rating = firstMatching(parts, /^\d(?:\.\d)?$/);
-        const reviewCountRaw = firstMatching(parts, /^\(?[\d,]+\)?(?:\s+reviews?)?$/i);
-        const phone = firstMatching(parts, /(?:\+?1[\s.-]?)?\(?\d{3}\)?[\s.-]\d{3}[\s.-]\d{4}/);
-        const hoursStatus = parts.find((part) => /^(open|closed|closes|opens)\b|^·\s*(opens|closes)\b/i.test(part)) ?? null;
-        const address = parts.find((part) => /\b[A-Z]{2}\s+\d{5}\b|\b(?:St|Street|Ave|Avenue|Rd|Road|Blvd|Drive|Dr)\b/i.test(part)) ?? null;
-        const directionsUrl = links.find((link) => /google\.[^/]+\/maps\/dir|\/dir\//i.test(link.href))?.href ?? `https://www.google.com/maps/dir/?api=1&destination=${encodeURIComponent([name, address].filter(Boolean).join(", ") || name)}`;
-        const excluded = normalizedSet([name, rating, reviewCountRaw, phone, hoursStatus, address, "Website", "Directions"]);
-        const category = parts.find((part) => {
-          const normalized = part.toLowerCase();
-          return !excluded.has(normalized) && !/^\d(?:\.\d)?$|^\(?[\d,]+\)?(?:\s+reviews?)?$/i.test(part) && !/(?:\+?1[\s.-]?)?\(?\d{3}\)?[\s.-]\d{3}[\s.-]\d{4}/.test(part) && !/\b[A-Z]{2}\s+\d{5}\b|\b(?:St|Street|Ave|Avenue|Rd|Road|Blvd|Drive|Dr)\b/i.test(part) && !/^(open|closed|closes|opens)\b|^·\s*(opens|closes)\b|directions|website|book online|sponsored|visit site|financing/i.test(part);
-        }) ?? null;
-        const { cid, cidDecimal } = cidFromUrl(placeUrl);
-        out.push({
-          position: out.length + 1,
-          name,
-          placeUrl,
-          cid,
-          cidDecimal,
-          rating,
-          reviewCount: reviewCountRaw ? reviewCountRaw.replace(/[()]/g, "") : null,
-          category,
-          address,
-          phone,
-          hoursStatus,
-          websiteUrl,
-          directionsUrl,
-          metadata: parts.slice(0, 20)
-        });
-      }
-      return out;
+      const visible = Array.from(document.querySelectorAll('dialog, [role="dialog"]')).find((dialog) => dialog.offsetParent !== null);
+      if (!visible) return { services: [], areasServed: [] };
+      const values = Array.from(visible.querySelectorAll("button")).map((button) => button.innerText.replace(/\s+/g, " ").trim()).filter(Boolean);
+      const servicesText = values.find((value) => /^services\s*:/i.test(value));
+      const services = servicesText ? [...new Set(servicesText.replace(/^services\s*:\s*/i, "").split(/\s*[,·]\s*/).map((value) => value.trim()).filter(Boolean))] : [];
+      const served = values.find((value) => /^serves\b/i.test(value));
+      const areasServed = served ? [served.replace(/^serves\s*/i, "").trim()].filter(Boolean) : [];
+      return { services, areasServed };
     });
   }
 };
@@ -14709,6 +14734,7 @@ async function searchCity(options, market) {
       gl: options.gl,
       hl: options.hl,
       maxResults: options.maxResultsPerCity,
+      includeServices: false,
       headless: options.headless,
       kernelApiKey: options.kernelApiKey,
       proxyMode: options.proxyMode,
@@ -16963,6 +16989,8 @@ import { setScopeTool } from "mcpscraper-memory-tools/tools/access/set-scope";
 import { inviteAccountTool } from "mcpscraper-memory-tools/tools/access/invite-account";
 import { getChatLinkTool } from "mcpscraper-memory-tools/tools/access/get-chat-link";
 import { revokeChatLinkTool } from "mcpscraper-memory-tools/tools/access/revoke-chat-link";
+import { getVaultAppLinkTool } from "mcpscraper-memory-tools/tools/access/get-vault-app-link";
+import { revokeVaultAppLinkTool } from "mcpscraper-memory-tools/tools/access/revoke-vault-app-link";
 import { provisionDefaultsTool as provisionDefaultsTool2 } from "mcpscraper-memory-tools/tools/vaults/provision-defaults";
 import { createSecureVaultTool } from "mcpscraper-memory-tools/tools/vaults/create-secure-vault";
 import { createScheduledActionTool } from "mcpscraper-memory-tools/tools/schedule/create-scheduled-action";
@@ -17004,6 +17032,14 @@ function registerBillingCutovers() {
     const result = await revokeChatLinkTool.execute({ ...input, apiKey }, {});
     return { content: [{ type: "text", text: JSON.stringify(result ?? {}) }], structuredContent: result ?? {}, isError: false };
   });
+  registerCutover("getVaultAppLinkTool", async (input, apiKey) => {
+    const result = await getVaultAppLinkTool.execute({ ...input, apiKey }, {});
+    return { content: [{ type: "text", text: JSON.stringify(result ?? {}) }], structuredContent: result ?? {}, isError: false };
+  });
+  registerCutover("revokeVaultAppLinkTool", async (input, apiKey) => {
+    const result = await revokeVaultAppLinkTool.execute({ ...input, apiKey }, {});
+    return { content: [{ type: "text", text: JSON.stringify(result ?? {}) }], structuredContent: result ?? {}, isError: false };
+  });
   registerCutover("provisionDefaultsTool", async (input, apiKey) => {
     const result = await provisionDefaultsTool2.execute({ ...input, apiKey }, {});
     return { content: [{ type: "text", text: JSON.stringify(result ?? {}) }], structuredContent: result ?? {}, isError: false };
@@ -20660,9 +20696,154 @@ loadChannels();
 </html>`;
 }
 
-// src/api/schedule-routes.ts
+// src/api/vault-routes.ts
 import { Hono as Hono22 } from "hono";
-var scheduleApp = new Hono22();
+var VAULTS = ["People", "Projects", "Tasks", "Communications"];
+var vaultApp = new Hono22();
+function validToken(token) {
+  return !!token && token.startsWith("mk_") && token.length > 20 && token.length < 256;
+}
+function isVault(value) {
+  return typeof value === "string" && VAULTS.includes(value);
+}
+function stringValue(value, max = 1e4) {
+  if (typeof value !== "string") return null;
+  const trimmed = value.trim();
+  return trimmed && trimmed.length <= max ? trimmed : null;
+}
+function propsValue(value) {
+  if (!value || typeof value !== "object" || Array.isArray(value)) return null;
+  const entries = Object.entries(value);
+  if (entries.length > 80) return null;
+  return Object.fromEntries(entries.filter(([key]) => key.length <= 96));
+}
+function cors(c) {
+  const origin = c.req.header("origin");
+  const configured = (process.env.VAULT_APP_ALLOWED_ORIGINS ?? "https://vault.mcpscraper.dev").split(",").map((value) => value.trim()).filter(Boolean);
+  const local = !!origin && /^http:\/\/localhost(?::\d+)?$/.test(origin);
+  if (origin && (configured.includes(origin) || local)) c.header("Access-Control-Allow-Origin", origin);
+  c.header("Vary", "Origin");
+  c.header("Access-Control-Allow-Methods", "GET, POST, OPTIONS");
+  c.header("Access-Control-Allow-Headers", "content-type");
+}
+vaultApp.use("/api/*", async (c, next) => {
+  cors(c);
+  if (c.req.method === "OPTIONS") return c.body(null, 204);
+  await next();
+});
+function workspaceRecord(vault, note) {
+  const props = note.props ?? {};
+  return {
+    id: `${vault}:${note.path}`,
+    vault,
+    path: note.path,
+    title: note.title,
+    content: note.content,
+    type: typeof props.type === "string" ? props.type : defaultType(vault),
+    status: typeof props.status === "string" ? props.status : defaultStatus(vault),
+    props,
+    revision: note.revision,
+    updatedAt: note.updatedAt,
+    capturedAt: note.capturedAt
+  };
+}
+function defaultType(vault) {
+  if (vault === "People") return "person";
+  if (vault === "Projects") return "project";
+  if (vault === "Tasks") return "task";
+  return "activity";
+}
+function defaultStatus(vault) {
+  if (vault === "People") return "active";
+  if (vault === "Projects") return "active";
+  if (vault === "Tasks") return "to_do";
+  return "triaged";
+}
+async function readVault(vault, token) {
+  const list = await memoryCall("listTool", { vault }, token);
+  if (!list.ok) throw new Error(list.error ?? `unable to read ${vault}`);
+  const noteRefs = (list.notes ?? []).slice(0, 200);
+  const notes = await Promise.all(
+    noteRefs.map(async ({ path: path5 }) => memoryCall("getTool", { vault, path: path5 }, token))
+  );
+  return notes.filter((result) => result.ok && !!result.note).map((result) => workspaceRecord(vault, result.note));
+}
+vaultApp.get("/api/workspace", async (c) => {
+  const token = c.req.query("token");
+  if (!validToken(token)) return c.json({ ok: false, error: "missing or malformed Vault App link" }, 401);
+  try {
+    const requested = c.req.query("vault");
+    const vaults = isVault(requested) ? [requested] : VAULTS;
+    const results = await Promise.all(vaults.map((vault) => readVault(vault, token)));
+    const records = results.flat().sort((a, b) => b.updatedAt.localeCompare(a.updatedAt));
+    return c.json({ ok: true, records, limited: results.some((result) => result.length >= 200) });
+  } catch (error) {
+    return c.json({ ok: false, error: error instanceof Error ? error.message : "unable to load vault workspace" }, 502);
+  }
+});
+vaultApp.post("/api/records", async (c) => {
+  const body = await c.req.json().catch(() => ({}));
+  const token = typeof body.token === "string" ? body.token : void 0;
+  if (!validToken(token)) return c.json({ ok: false, error: "missing or malformed Vault App link" }, 401);
+  const vault = body.vault;
+  const path5 = stringValue(body.path, 320);
+  const title = stringValue(body.title, 320);
+  const content = typeof body.content === "string" && body.content.length <= 1e5 ? body.content : "";
+  const props = propsValue(body.props) ?? {};
+  if (!isVault(vault) || !path5 || !title) return c.json({ ok: false, error: "vault, path, and title are required" }, 400);
+  const baseRevision = typeof body.baseRevision === "number" && Number.isInteger(body.baseRevision) ? body.baseRevision : void 0;
+  const result = await memoryCall("putTool", { vault, path: path5, title, content: content || title, props, ...baseRevision === void 0 ? {} : { baseRevision } }, token);
+  return c.json(result, result.ok ? 200 : result.conflict ? 409 : 502);
+});
+vaultApp.post("/api/records/transition", async (c) => {
+  const body = await c.req.json().catch(() => ({}));
+  const token = typeof body.token === "string" ? body.token : void 0;
+  if (!validToken(token)) return c.json({ ok: false, error: "missing or malformed Vault App link" }, 401);
+  const vault = typeof body.vault === "string" ? body.vault : "";
+  const path5 = typeof body.path === "string" ? body.path : "";
+  const status = stringValue(body.status, 80);
+  const baseRevision = typeof body.baseRevision === "number" && Number.isInteger(body.baseRevision) ? body.baseRevision : void 0;
+  if (!isVault(vault) || !path5 || !status) return c.json({ ok: false, error: "valid vault, path, and status are required" }, 400);
+  const current = await memoryCall("getTool", { vault, path: path5 }, token);
+  if (!current.ok || !current.note) return c.json(current, current.ok ? 404 : 502);
+  const result = await memoryCall("putTool", {
+    vault,
+    path: path5,
+    title: current.note.title,
+    content: current.note.content,
+    props: { ...current.note.props ?? {}, status },
+    baseRevision: baseRevision ?? current.note.revision
+  }, token);
+  return c.json(result, result.ok ? 200 : result.conflict ? 409 : 502);
+});
+vaultApp.post("/api/activity", async (c) => {
+  const body = await c.req.json().catch(() => ({}));
+  const token = typeof body.token === "string" ? body.token : void 0;
+  if (!validToken(token)) return c.json({ ok: false, error: "missing or malformed Vault App link" }, 401);
+  const title = stringValue(body.title, 320);
+  const content = typeof body.content === "string" && body.content.length <= 2e4 ? body.content : "";
+  const props = propsValue(body.props) ?? {};
+  if (!title) return c.json({ ok: false, error: "title is required" }, 400);
+  const stamp = (/* @__PURE__ */ new Date()).toISOString().replace(/[:.]/g, "-");
+  const result = await memoryCall("putTool", {
+    vault: "Communications",
+    path: `Inbox/activity-${stamp}`,
+    title,
+    content: content || title,
+    props: { status: "triaged", type: "decision", occurred_at: (/* @__PURE__ */ new Date()).toISOString(), ...props }
+  }, token);
+  return c.json(result, result.ok ? 200 : 502);
+});
+vaultApp.get("/:token", (c) => {
+  const token = c.req.param("token");
+  if (!validToken(token)) return c.html('<!doctype html><html><body style="font-family:system-ui;padding:40px"><h2>This Vault App link is not valid.</h2><p>Ask the AI to run get-vault-app-link again, or revoke-vault-app-link and issue a replacement.</p></body></html>', 401);
+  const base = (process.env.VAULT_APP_BASE_URL ?? "https://vault.mcpscraper.dev").replace(/\/$/, "");
+  return c.redirect(`${base}/vault/${encodeURIComponent(token)}`, 302);
+});
+
+// src/api/schedule-routes.ts
+import { Hono as Hono23 } from "hono";
+var scheduleApp = new Hono23();
 function tokenOk2(token) {
   return !!token && token.startsWith("mk_") && token.length > 10;
 }
@@ -20993,7 +21174,7 @@ loadActions();
 }
 
 // src/api/resend-inbound-routes.ts
-import { Hono as Hono23 } from "hono";
+import { Hono as Hono24 } from "hono";
 import { Resend } from "resend";
 var SUPPORT_ADDRESS = "support@mcpscraper.dev";
 var DMARC_ADDRESS = "dmarc@mcpscraper.dev";
@@ -21169,7 +21350,7 @@ async function forwardReceivedEmail(resend, email, destination, label) {
   if (sent.error) throw new Error(`forward failed: ${sent.error.message}`);
   return sent.data.id;
 }
-var resendInboundApp = new Hono23();
+var resendInboundApp = new Hono24();
 resendInboundApp.post("/webhooks", async (c) => {
   const apiKey = process.env.RESEND_API_KEY?.trim();
   const webhookSecret = process.env.RESEND_WEBHOOK_SECRET?.trim();
@@ -24754,7 +24935,7 @@ var requirePaidSchedulingTier = createMiddleware3(async (c, next) => {
   if (!hasPaidSchedulingAccess(user)) return c.json(paidPlanRequiredPayload(), 403);
   return next();
 });
-var app = new Hono24();
+var app = new Hono25();
 var STRIPE_API_VERSION2 = "2026-02-25.clover";
 function requireStripeSecret() {
   const secret2 = process.env.STRIPE_SECRET_KEY?.trim();
@@ -26978,7 +27159,7 @@ app.get("/cron/tick", async (c) => {
   if (!process.env.CRON_SECRET || secret2 !== `Bearer ${process.env.CRON_SECRET}`) {
     return c.json({ error: "Unauthorized" }, 401);
   }
-  const { drainQueue } = await import("./worker-E3KHQH2F.js");
+  const { drainQueue } = await import("./worker-RYP5EOV5.js");
   const budget = { maxJobs: 10, deadlineMs: Date.now() + 28e4 };
   const workflowDispatchResult = await dispatchDueWorkflowSchedules(`${new URL(c.req.url).protocol}//${new URL(c.req.url).host}`);
   const [results, sweepResult, reapResult, expiredResult, blobCleanup, connectedDataArtifactCleanup] = await Promise.all([
@@ -27049,6 +27230,7 @@ app.route("/serp-intelligence", serpIntelligenceApp);
 app.route("/mcp", mcpApp);
 app.route("/agent", buildBrowserAgentRoutes());
 app.route("/chat", chatApp);
+app.route("/vault", vaultApp);
 app.route("/schedule", scheduleApp);
 app.route("/resend", resendInboundApp);
 app.get("/console", (c) => c.html(renderConsoleHtml()));
@@ -27180,4 +27362,4 @@ app.get("/blog/:slug/", (c) => {
 export {
   app
 };
-//# sourceMappingURL=server-GN34MGAE.js.map
+//# sourceMappingURL=server-QQB4BNAB.js.map