mcp-scraper 0.32.3 → 0.33.1

package/README.md

@@ -88,7 +88,7 @@ Build the branded one-click bundle:
 npm run build:mcpb
 ```
 
-The generated bundle is written to `build/mcpb/mcp-scraper-<version>.mcpb` and copied to `public/downloads/` for the hosted download. The current public bundle is `https://mcpscraper.dev/downloads/mcp-scraper.mcpb` (`0.32.3`, SHA-256 `1e44cc581a788e799b5e0eb7e909178714a50030d86bbdc6c2b119d84b04fb97`). Install it by opening or dragging it into Claude Desktop. Claude displays the `MCP Scraper` install card, icon, and API-key configuration field from the bundle manifest.
+The generated bundle is written to `build/mcpb/mcp-scraper-<version>.mcpb` and copied to `public/downloads/` for the hosted download. The current public bundle is `https://mcpscraper.dev/downloads/mcp-scraper.mcpb` (`0.33.1`, SHA-256 `c9122255263b899e7ac95825d0212c54f63059be0370dabd67aea57af82f5754`). Install it by opening or dragging it into Claude Desktop. Claude displays the `MCP Scraper` install card, icon, and API-key configuration field from the bundle manifest.
 
 The MCPB install exposes every tool — web-intelligence plus all `browser_*` tools — through the one `mcp-scraper` server.
 

package/dist/bin/api-server.cjs

@@ -15945,7 +15945,7 @@ var init_server_schemas = __esm({
       gl: import_zod15.z.string().optional(),
       hl: import_zod15.z.string().optional(),
       device: import_zod15.z.enum(["desktop", "mobile"]).optional(),
-      proxyMode: import_zod15.z.enum(["location", "configured", "none"]).optional(),
+      proxyMode: import_zod15.z.enum(["configured", "none"]).optional(),
       proxyZip: import_zod15.z.string().regex(/^\d{5}$/).optional(),
       debug: import_zod15.z.boolean().optional(),
       serpOnly: import_zod15.z.boolean().optional(),
@@ -19253,13 +19253,11 @@ var init_video_routes = __esm({
 function buildGoogleLocalResultsUrl(input) {
   const query = [input.query, input.location].filter(Boolean).join(" ");
   const params = new URLSearchParams({ q: query, udm: "1", gl: input.gl, hl: input.hl, pws: "0" });
-  if (input.location) params.set("uule", encodeUule(normalizeLocation(input.location)));
   return `https://www.google.com/search?${params.toString()}`;
 }
 function buildGoogleOrganicSearchUrl(input) {
   const query = [input.query, input.location].filter(Boolean).join(" ");
   const params = new URLSearchParams({ q: query, 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()}`;
 }
 function mapsIdentifiersFromUrl(url) {
@@ -19284,7 +19282,6 @@ var init_MapsSearchExtractor = __esm({
     "use strict";
     init_errors();
     init_BrowserDriver();
-    init_uule();
     LOCAL_RESULTS_PAGE_LIMIT = 8;
     LOCAL_RESULTS_WAIT_MS = 1e3;
     MAPS_REDIRECT_BASE = "https://serp-redirector.vercel.app";
@@ -20092,7 +20089,9 @@ var init_MapsExtractor = __esm({
 
 // src/maps-search-rotation.ts
 function mapsSearchMaxAttemptsForProxyMode(proxyMode) {
-  return proxyMode === "location" ? MAPS_SEARCH_LOCATION_PROXY_MAX_ATTEMPTS : MAPS_SEARCH_MAX_ATTEMPTS;
+  if (proxyMode === "location") return MAPS_SEARCH_LOCATION_PROXY_MAX_ATTEMPTS;
+  if (proxyMode === "none") return MAPS_SEARCH_NONE_MAX_ATTEMPTS;
+  return MAPS_SEARCH_MAX_ATTEMPTS;
 }
 function mapsSearchErrorMessage(err) {
   return sanitizeVendorName(err instanceof Error ? err.message : String(err));
@@ -20163,6 +20162,7 @@ async function cleanupDisposableProxy(kernelApiKey, proxyId, eventName) {
   }
 }
 async function runMapsSearchWithRotation(options) {
+  if (options.proxyMode === "location") options = { ...options, proxyMode: "none" };
   const attempts = [];
   const maxAttempts = mapsSearchMaxAttemptsForProxyMode(options.proxyMode) + 1;
   const started = Date.now();
@@ -20191,8 +20191,8 @@ async function runMapsSearchWithRotation(options) {
         proxyZip,
         kernelProxyId: options.proxyMode === "none" ? void 0 : resolution2.kernelProxyId,
         kernelProxyResolution: resolution2.resolution,
-        serpRedirect: options.proxyMode === "none" && attemptIndex >= 1 && attemptIndex < maxAttempts - 1,
-        forceDirectEgress: options.proxyMode === "none" && attemptIndex >= 1
+        serpRedirect: options.proxyMode === "none" && attemptIndex >= 2 && attemptIndex < 6,
+        forceDirectEgress: options.proxyMode === "none" && attemptIndex >= 4
       };
       const isHeadfulEscalation = attemptIndex === maxAttempts - 1;
       const baseCtx = currentCostContext();
@@ -20267,7 +20267,7 @@ async function runMapsSearchWithRotation(options) {
     true
   );
 }
-var MAPS_SEARCH_MAX_ATTEMPTS, MAPS_SEARCH_LOCATION_PROXY_MAX_ATTEMPTS, MapsSearchRotationError;
+var MAPS_SEARCH_MAX_ATTEMPTS, MAPS_SEARCH_NONE_MAX_ATTEMPTS, MAPS_SEARCH_LOCATION_PROXY_MAX_ATTEMPTS, MapsSearchRotationError;
 var init_maps_search_rotation = __esm({
   "src/maps-search-rotation.ts"() {
     "use strict";
@@ -20277,6 +20277,7 @@ var init_maps_search_rotation = __esm({
     init_kernel_proxy_resolver();
     init_cost_context();
     MAPS_SEARCH_MAX_ATTEMPTS = 3;
+    MAPS_SEARCH_NONE_MAX_ATTEMPTS = 7;
     MAPS_SEARCH_LOCATION_PROXY_MAX_ATTEMPTS = 5;
     MapsSearchRotationError = class extends Error {
       constructor(message, attempts, retryable, errorCode = "maps_search_failed") {
@@ -24138,7 +24139,7 @@ var init_directory_workflow = __esm({
       saveCsv: import_zod26.z.boolean().default(true),
       gl: import_zod26.z.string().length(2).default("us"),
       hl: import_zod26.z.string().length(2).default("en"),
-      proxyMode: import_zod26.z.enum(["location", "configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE),
+      proxyMode: import_zod26.z.enum(["configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE),
       proxyZip: import_zod26.z.string().regex(/^\d{5}$/).optional(),
       debug: import_zod26.z.boolean().default(false),
       headless: import_zod26.z.boolean().default(true),
@@ -24794,7 +24795,7 @@ var init_directory = __esm({
       maxCities: import_zod28.z.number().int().min(1).max(100).default(25),
       maxResultsPerCity: import_zod28.z.number().int().min(1).max(50).default(20),
       concurrency: import_zod28.z.number().int().min(1).max(5).default(5),
-      proxyMode: import_zod28.z.enum(["location", "configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE),
+      proxyMode: import_zod28.z.enum(["configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE),
       saveCsv: import_zod28.z.boolean().default(true)
     });
     DIRECTORY_CSV_HEADERS = [
@@ -24949,7 +24950,7 @@ var init_get_leads = __esm({
       enrichWebsites: import_zod29.z.boolean().default(true).describe("Visit each business website (home + contact pages) to harvest email and social links. Uses the proxy/browser-backed extractor so blocked sites still resolve."),
       hydrateReviewCounts: import_zod29.z.boolean().default(true).describe("Deep-dive each profile to confirm the review count and booking URL that the Maps search list omits."),
       concurrency: import_zod29.z.number().int().min(1).max(4).default(3).describe("How many businesses to enrich in parallel. Keep low to respect per-account concurrency limits."),
-      proxyMode: import_zod29.z.enum(["location", "configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE).describe("Proxy targeting for the Maps search. location targets the market; none is for local debugging.")
+      proxyMode: import_zod29.z.enum(["configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE).describe("Proxy behavior for the Maps search. Leave unset for clean egress; country/region localization comes from gl/hl plus the city or region in the query.")
     });
     LEADS_CSV_HEADERS = [
       "position",
@@ -25168,7 +25169,7 @@ var init_local_competitive_audit = __esm({
       hydrateTop: import_zod30.z.number().int().min(0).max(10).default(5),
       maxReviews: import_zod30.z.number().int().min(0).max(500).default(50),
       concurrency: import_zod30.z.number().int().min(1).max(5).default(5),
-      proxyMode: import_zod30.z.enum(["location", "configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE),
+      proxyMode: import_zod30.z.enum(["configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE),
       returnPartial: import_zod30.z.boolean().default(true)
     });
     localCompetitiveAuditWorkflowDefinition = {
@@ -25621,7 +25622,7 @@ var init_comparison_briefs = __esm({
     init_directory();
     init_seo_workflow_utils();
     init_schemas3();
-    ProxyModeSchema = import_zod31.z.enum(["location", "configured", "none"]);
+    ProxyModeSchema = import_zod31.z.enum(["configured", "none"]);
     MapComparisonInputSchema = import_zod31.z.object({
       query: import_zod31.z.string().min(1),
       location: import_zod31.z.string().optional(),
@@ -29814,7 +29815,7 @@ var PACKAGE_VERSION;
 var init_version = __esm({
   "src/version.ts"() {
     "use strict";
-    PACKAGE_VERSION = "0.32.3";
+    PACKAGE_VERSION = "0.33.1";
   }
 });
 
@@ -30278,7 +30279,7 @@ var init_mcp_tool_schemas = __esm({
       gl: import_zod35.z.string().length(2).default("us").describe("Google country code inferred from location or user language."),
       hl: import_zod35.z.string().default("en").describe("Google interface/content language inferred from the user request."),
       device: import_zod35.z.enum(["desktop", "mobile"]).default("desktop").describe("SERP device context. Use mobile only for mobile rankings."),
-      proxyMode: import_zod35.z.enum(["location", "configured", "none"]).default(DEFAULT_PROXY_MODE).describe('Leave unset (clean egress). Do NOT set "location" just because a city was named \u2014 that comes from city-in-query wording. "location" forces residential geo-IP for rank-tracking fidelity, is frequently CAPTCHA-blocked, and accepts failures.'),
+      proxyMode: import_zod35.z.enum(["configured", "none"]).default(DEFAULT_PROXY_MODE).describe("Leave unset for clean egress (the default). Country/region localization comes from gl/hl plus the city or region in the query \u2014 not from a proxy."),
       proxyZip: import_zod35.z.string().regex(/^\d{5}$/).optional().describe('US ZIP for residential geo-IP targeting. Only meaningful with proxyMode "location".'),
       debug: import_zod35.z.boolean().default(false).describe("Include sanitized diagnostics for debugging localization, CAPTCHA, or proxy behavior.")
     };
@@ -30447,7 +30448,7 @@ var init_mcp_tool_schemas = __esm({
       hl: import_zod35.z.string().length(2).default("en").describe("Language inferred from user request."),
       maxResults: import_zod35.z.number().int().min(1).max(50).default(10).describe("Number of candidates to return. Default 10, maximum 50."),
       includeServices: import_zod35.z.boolean().default(false).describe("Open each returned business profile to include its configured services and areas served when available. Adds a page visit per business; does not collect review cards."),
-      proxyMode: import_zod35.z.enum(["location", "configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE).describe("Leave unset for the default route (stealth browser on the managed ISP proxy, retried on a fresh session when Google blocks). Localization comes from the city in the query plus gl/hl. location forces an explicit residential proxy \u2014 not recommended for Google."),
+      proxyMode: import_zod35.z.enum(["configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE).describe("Leave unset for the default route (stealth browser on the managed ISP proxy, retried on a fresh session when Google blocks). Country/region localization comes from the city or region in the query plus gl/hl \u2014 not from a proxy."),
       proxyZip: import_zod35.z.string().regex(/^\d{5}$/).optional().describe("Optional US ZIP override, only used when proxyMode is location."),
       debug: import_zod35.z.boolean().default(false).describe("Include sanitized browser/proxy diagnostics.")
     };
@@ -30462,7 +30463,7 @@ var init_mcp_tool_schemas = __esm({
       includeZipGroups: import_zod35.z.boolean().default(true).describe("Attach ZIP groups from a configured US ZIPS CSV when available (MCP_SCRAPER_USZIPS_CSV_PATH or usZipsCsvPath)."),
       usZipsCsvPath: import_zod35.z.string().optional().describe("Local/test-only path to a US ZIPS CSV (state_abbr, zipcode, county, city columns). Deployed APIs should use MCP_SCRAPER_USZIPS_CSV_PATH instead. For ZIP enrichment, set MCP_SCRAPER_USZIPS_CSV_PATH on the server, or pass this in local/test mode."),
       saveCsv: import_zod35.z.boolean().default(true).describe("Save a directory-ready CSV of results to the MCP Scraper output directory and return its path."),
-      proxyMode: import_zod35.z.enum(["location", "configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE).describe("Proxy behavior per city search. Leave unset for the default route (stealth browser on the managed ISP proxy, retried fresh on a Google block). location forces an explicit residential proxy \u2014 not recommended for Google."),
+      proxyMode: import_zod35.z.enum(["configured", "none"]).default(DEFAULT_MAPS_PROXY_MODE).describe("Proxy behavior per city search. Leave unset for the default route (stealth browser on the managed ISP proxy, retried fresh on a Google block). Country/region localization comes from the city or region in the query plus gl/hl \u2014 not from a proxy."),
       proxyZip: import_zod35.z.string().regex(/^\d{5}$/).optional().describe("Optional ZIP override for proxy targeting; normally omitted."),
       debug: import_zod35.z.boolean().default(false).describe("Include sanitized browser/proxy diagnostics.")
     };
@@ -31302,7 +31303,7 @@ var init_mcp_tool_schemas = __esm({
       gl: import_zod35.z.string().length(2).default("us").describe("Google country code inferred from location or user language."),
       hl: import_zod35.z.string().default("en").describe("Google interface/content language inferred from user request."),
       device: import_zod35.z.enum(["desktop", "mobile"]).default("desktop").describe("SERP device context. Use mobile only for mobile rankings."),
-      proxyMode: import_zod35.z.enum(["location", "configured", "none"]).default(DEFAULT_PROXY_MODE).describe('Leave unset (clean egress). Do NOT set "location" just because a city was named \u2014 that comes from city-in-query wording. "location" forces residential geo-IP for rank-tracking fidelity, is frequently CAPTCHA-blocked, and accepts failures.'),
+      proxyMode: import_zod35.z.enum(["configured", "none"]).default(DEFAULT_PROXY_MODE).describe("Leave unset for clean egress (the default). Country/region localization comes from gl/hl plus the city or region in the query \u2014 not from a proxy."),
       proxyZip: import_zod35.z.string().regex(/^\d{5}$/).optional().describe('US ZIP for residential geo-IP targeting. Only meaningful with proxyMode "location".'),
       debug: import_zod35.z.boolean().default(false).describe("Include sanitized diagnostics for debugging localization, CAPTCHA, or proxy behavior."),
       pages: import_zod35.z.number().int().min(1).max(2).default(1).describe("Number of result pages to fetch (1\u20132).")
@@ -31313,7 +31314,7 @@ var init_mcp_tool_schemas = __esm({
       gl: import_zod35.z.string().length(2).default("us").describe("Google country code inferred from the requested market."),
       hl: import_zod35.z.string().default("en").describe("Google interface/content language inferred from the user request."),
       device: import_zod35.z.enum(["desktop", "mobile"]).default("desktop").describe("SERP device context. Use mobile only for mobile rankings/evidence."),
-      proxyMode: import_zod35.z.enum(["location", "configured", "none"]).default(DEFAULT_PROXY_MODE).describe('Leave unset (clean egress). Do NOT set "location" just because a city was named \u2014 that comes from city-in-query wording. "location" forces residential geo-IP, is frequently CAPTCHA-blocked, and accepts failures.'),
+      proxyMode: import_zod35.z.enum(["configured", "none"]).default(DEFAULT_PROXY_MODE).describe("Leave unset for clean egress (the default). Country/region localization comes from gl/hl plus the city or region in the query \u2014 not from a proxy."),
       proxyZip: import_zod35.z.string().regex(/^\d{5}$/).optional().describe('US ZIP for residential geo-IP targeting. Only meaningful with proxyMode "location".'),
       pages: import_zod35.z.number().int().min(1).max(2).default(1).describe("Google result pages to capture. Use 2 only for deeper ranking evidence."),
       debug: import_zod35.z.boolean().default(false).describe("Include sanitized browser/proxy/location diagnostics."),