mcpscraper-cli 0.39.1 → 0.40.0

package/README.md

@@ -47,6 +47,11 @@ Integration connections are included with an active Starter plan or higher and h
 
 Current Google search pricing is 60 Credits per SERP search and 400 Credits plus 10 Credits per returned question for PAA. One optional concurrency pack adds two browser slots for $5/month; quantity n adds 2n browser slots.
 
+Use `mcpscraper tools call harvest_paa_start --args '{"query":"...","pages":2,"idempotencyKey":"..."}' --json`
+to request a second organic-results page, then poll the returned `jobId` with
+`harvest_paa_status`. One page is the default, and the second page never creates
+a second PAA graph. Pagination can be null for older saved jobs.
+
 When a tool returns inline MCP media, the CLI writes each block to a private local file and prints its path instead of dumping base64. Use `--media-dir <path>` to choose the destination; the default is `~/Downloads/mcp-scraper`.
 
 ### `mcpscraper search <query>`

package/dist/bin/cli.js

@@ -10,7 +10,7 @@ import { randomUUID } from "crypto";
 import { ScraperClient, ScraperApiError } from "mcpscraper-sdk";
 
 // src/version.ts
-var CLI_VERSION = "0.39.1";
+var CLI_VERSION = "0.40.0";
 
 // src/generated-tools.ts
 var MCP_TOOL_CATALOG = [
@@ -18760,10 +18760,17 @@ var MCP_TOOL_CATALOG = [
     "name": "harvest_paa",
     "category": "search",
     "title": "Google PAA + SERP Harvest",
-    "description": "Expand one Google People Also Ask SERP into questions, answers, every preserved source, AI Overview evidence, ranking URLs, and entity IDs. maxQuestions is a target count, not traversal depth. Results distinguish target_reached, proven frontier_exhausted, interruption, and recovery_exhausted; a failed click or browser timeout is never reported as exhaustion. This compatibility tool waits; use harvest_paa_start plus harvest_paa_status for long runs. Optional SERP modules require their include flags. Use gl and location for regional context. Costs 400 Credits per harvest plus 10 Credits per question actually returned; unused hold is refunded. After a timeout or unknown response, reuse the same idempotencyKey. Call credits_info for current pricing and balance.",
+    "description": "Expand one Google People Also Ask SERP into questions, answers, every preserved source, AI Overview evidence, ranking URLs, and entity IDs. Set pages to 2 to add the second organic-results page when available; PAA is still expanded once on the preserved first page, and pagination reports what was captured. maxQuestions is a target count, not traversal depth. Results distinguish target_reached, proven frontier_exhausted, interruption, and recovery_exhausted; a failed click or browser timeout is never reported as exhaustion. This compatibility tool waits; use harvest_paa_start plus harvest_paa_status for long runs. Optional SERP modules require their include flags. Use gl and location for regional context. Costs 400 Credits per harvest plus 10 Credits per question actually returned; unused hold is refunded. After a timeout or unknown response, reuse the same idempotencyKey. Call credits_info for current pricing and balance.",
     "inputSchema": {
       "type": "object",
       "properties": {
+        "pages": {
+          "default": 1,
+          "description": "Organic result pages to capture. Default 1, maximum 2. Page 2 is captured when available before harvesting PAA on the original first page; it does not add a second PAA graph. Pagination output reports the pages actually captured.",
+          "type": "integer",
+          "minimum": 1,
+          "maximum": 2
+        },
         "query": {
           "type": "string",
           "minLength": 1,
@@ -18860,10 +18867,17 @@ var MCP_TOOL_CATALOG = [
     "name": "harvest_paa_start",
     "category": "other",
     "title": "Start Durable Google PAA Harvest",
-    "description": "Start a durable Google People Also Ask harvest and return its job receipt. maxQuestions is the requested target count. The job automatically resumes an interrupted serverless worker under the same jobId, idempotency key, and billing hold while preserving checkpoints. Keep one idempotencyKey after a timeout, unknown response, or in-progress reply; replaying it recovers the existing job without a duplicate charge. Poll jobId with harvest_paa_status. Costs 400 Credits plus 10 per retained question; unused hold is refunded.",
+    "description": "Start a durable Google People Also Ask harvest and return its job receipt. Set pages to 2 to add the second organic-results page when available; PAA is still expanded once on the preserved first page, and pagination survives checkpoint recovery. maxQuestions is the requested target count. The job automatically resumes an interrupted serverless worker under the same jobId, idempotency key, and billing hold while preserving checkpoints. Keep one idempotencyKey after a timeout, unknown response, or in-progress reply; replaying it recovers the existing job without a duplicate charge. Poll jobId with harvest_paa_status. Costs 400 Credits plus 10 per retained question; unused hold is refunded.",
     "inputSchema": {
       "type": "object",
       "properties": {
+        "pages": {
+          "default": 1,
+          "description": "Organic result pages to capture. Default 1, maximum 2. Page 2 is captured when available before harvesting PAA on the original first page; it does not add a second PAA graph. Pagination output reports the pages actually captured.",
+          "type": "integer",
+          "minimum": 1,
+          "maximum": 2
+        },
         "query": {
           "type": "string",
           "minLength": 1,
@@ -18961,7 +18975,7 @@ var MCP_TOOL_CATALOG = [
     "name": "harvest_paa_status",
     "category": "other",
     "title": "Check Durable Google PAA Harvest",
-    "description": "Poll an owner-scoped harvest_paa_start job. Returns the job/operation/task correlation ID, saved progress, automatic-recovery count, target and discovery status, material completeness, recent per-control interaction outcomes with 0.7/1.0/1.4-second confirmation telemetry, provider-session-correlated attempts, terminal rows, and billing. frontier_exhausted means every observed eligible PAA control was processed plus three healthy no-growth confirmations; interruption never means exhaustion. Polling never starts or bills another run.",
+    "description": "Poll an owner-scoped harvest_paa_start job. Returns the job/operation/task correlation ID, saved progress, automatic-recovery count, target and discovery status, material completeness, bounded organic-page pagination, recent per-control interaction outcomes with 0.7/1.0/1.4-second confirmation telemetry, provider-session-correlated attempts, terminal rows, and billing. Page-two outcomes describe organic results only; the PAA graph always comes from the preserved first page. frontier_exhausted means every observed eligible PAA control was processed plus three healthy no-growth confirmations; interruption never means exhaustion. Polling never starts or bills another run.",
     "inputSchema": {
       "type": "object",
       "properties": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mcpscraper-cli",
-  "version": "0.39.1",
+  "version": "0.40.0",
   "description": "Command-line interface for mcpscraper.dev and memory.mcpscraper.dev",
   "license": "MIT",
   "type": "module",
@@ -19,7 +19,7 @@
   },
   "dependencies": {
     "commander": "^12.0.0",
-    "mcpscraper-sdk": "^0.40.1"
+    "mcpscraper-sdk": "^0.41.0"
   },
   "devDependencies": {
     "@types/node": "^22.0.0"