context.dev 1.16.0 → 1.18.0

data/lib/context_dev/resources/web.rb

@@ -70,7 +70,7 @@ module ContextDev
       #
       # Capture a screenshot of a website.
       #
-      # @overload screenshot(direct_url: nil, domain: nil, full_screenshot: nil, max_age_ms: nil, page: nil, prioritize: nil, viewport: nil, request_options: {})
+      # @overload screenshot(direct_url: nil, domain: nil, full_screenshot: nil, max_age_ms: nil, page: nil, timeout_ms: nil, viewport: nil, wait_for_ms: nil, request_options: {})
       #
       # @param direct_url [String] A specific URL to screenshot directly, bypassing domain resolution (e.g., 'https
       #
@@ -82,10 +82,12 @@ module ContextDev
       #
       # @param page [Symbol, ContextDev::Models::WebScreenshotParams::Page] Optional parameter to specify which page type to screenshot. If provided, the sy
       #
-      # @param prioritize [Symbol, ContextDev::Models::WebScreenshotParams::Prioritize] Optional parameter to prioritize screenshot capture. If 'speed', optimizes for f
+      # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
       #
       # @param viewport [ContextDev::Models::WebScreenshotParams::Viewport] Optional browser viewport dimensions for the screenshot. Defaults to 1920x1080.
       #
+      # @param wait_for_ms [Integer] Optional browser wait time in milliseconds after initial page load before taking
+      #
       # @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}, nil]
       #
       # @return [ContextDev::Models::WebScreenshotResponse]
@@ -100,7 +102,9 @@ module ContextDev
           query: query.transform_keys(
             direct_url: "directUrl",
             full_screenshot: "fullScreenshot",
-            max_age_ms: "maxAgeMs"
+            max_age_ms: "maxAgeMs",
+            timeout_ms: "timeoutMS",
+            wait_for_ms: "waitForMs"
           ),
           model: ContextDev::Models::WebScreenshotResponse,
           options: options
@@ -113,7 +117,7 @@ module ContextDev
       # Performs a crawl starting from a given URL, extracts page content as Markdown,
       # and returns results for all crawled pages.
       #
-      # @overload web_crawl_md(url:, follow_subdomains: nil, include_frames: nil, include_images: nil, include_links: nil, max_age_ms: nil, max_depth: nil, max_pages: nil, parse_pdf: nil, shorten_base64_images: nil, url_regex: nil, use_main_content_only: nil, request_options: {})
+      # @overload web_crawl_md(url:, follow_subdomains: nil, include_frames: nil, include_images: nil, include_links: nil, max_age_ms: nil, max_depth: nil, max_pages: nil, pdf: nil, shorten_base64_images: nil, stop_after_ms: nil, timeout_ms: nil, url_regex: nil, use_main_content_only: nil, wait_for_ms: nil, request_options: {})
       #
       # @param url [String] The starting URL for the crawl (must include http:// or https:// protocol)
       #
@@ -131,14 +135,20 @@ module ContextDev
       #
       # @param max_pages [Integer] Maximum number of pages to crawl. Hard cap: 500.
       #
-      # @param parse_pdf [Boolean] When true (default), PDF pages are fetched and their text layer is extracted and
+      # @param pdf [ContextDev::Models::WebWebCrawlMdParams::Pdf] PDF parsing controls. Use start/end to limit text extraction and OCR to an inclu
       #
       # @param shorten_base64_images [Boolean] Truncate base64-encoded image data in the Markdown output
       #
+      # @param stop_after_ms [Integer] Soft time budget for the crawl in milliseconds. After each scrape, the crawler c
+      #
+      # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
+      #
       # @param url_regex [String] Regex pattern. Only URLs matching this pattern will be followed and scraped.
       #
       # @param use_main_content_only [Boolean] Extract only the main content, stripping headers, footers, sidebars, and navigat
       #
+      # @param wait_for_ms [Integer] Optional browser wait time in milliseconds after initial page load for each craw
+      #
       # @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}, nil]
       #
       # @return [ContextDev::Models::WebWebCrawlMdResponse]
@@ -160,7 +170,7 @@ module ContextDev
       #
       # Scrapes the given URL and returns the raw HTML content of the page.
       #
-      # @overload web_scrape_html(url:, include_frames: nil, max_age_ms: nil, parse_pdf: nil, request_options: {})
+      # @overload web_scrape_html(url:, include_frames: nil, max_age_ms: nil, pdf: nil, timeout_ms: nil, wait_for_ms: nil, request_options: {})
       #
       # @param url [String] Full URL to scrape (must include http:// or https:// protocol)
       #
@@ -168,7 +178,11 @@ module ContextDev
       #
       # @param max_age_ms [Integer] Return a cached result if a prior scrape for the same parameters exists and is y
       #
-      # @param parse_pdf [Boolean] When true (default), PDF URLs are fetched and their text layer is extracted and
+      # @param pdf [ContextDev::Models::WebWebScrapeHTMLParams::Pdf] PDF parsing controls. Use start/end to limit text extraction and OCR to an inclu
+      #
+      # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
+      #
+      # @param wait_for_ms [Integer] Optional browser wait time in milliseconds after initial page load. Min: 0. Max:
       #
       # @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}, nil]
       #
@@ -184,7 +198,8 @@ module ContextDev
           query: query.transform_keys(
             include_frames: "includeFrames",
             max_age_ms: "maxAgeMs",
-            parse_pdf: "parsePDF"
+            timeout_ms: "timeoutMS",
+            wait_for_ms: "waitForMs"
           ),
           model: ContextDev::Models::WebWebScrapeHTMLResponse,
           options: options
@@ -199,7 +214,7 @@ module ContextDev
       # embeds. The base request costs 1 credit; enrichment costs 1 credit per returned
       # image.
       #
-      # @overload web_scrape_images(url:, enrichment: nil, max_age_ms: nil, request_options: {})
+      # @overload web_scrape_images(url:, enrichment: nil, max_age_ms: nil, timeout_ms: nil, wait_for_ms: nil, request_options: {})
       #
       # @param url [String] Page URL to inspect. Must include http:// or https://.
       #
@@ -207,6 +222,10 @@ module ContextDev
       #
       # @param max_age_ms [Integer] Reuse a cached result this many milliseconds old or newer. Default: 86400000 (1
       #
+      # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
+      #
+      # @param wait_for_ms [Integer] Optional browser wait time in milliseconds after initial page load before collec
+      #
       # @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}, nil]
       #
       # @return [ContextDev::Models::WebWebScrapeImagesResponse]
@@ -218,7 +237,11 @@ module ContextDev
         @client.request(
           method: :get,
           path: "web/scrape/images",
-          query: query.transform_keys(max_age_ms: "maxAgeMs"),
+          query: query.transform_keys(
+            max_age_ms: "maxAgeMs",
+            timeout_ms: "timeoutMS",
+            wait_for_ms: "waitForMs"
+          ),
           model: ContextDev::Models::WebWebScrapeImagesResponse,
           options: options
         )
@@ -229,7 +252,7 @@ module ContextDev
       #
       # Scrapes the given URL into LLM usable Markdown.
       #
-      # @overload web_scrape_md(url:, include_frames: nil, include_images: nil, include_links: nil, max_age_ms: nil, parse_pdf: nil, shorten_base64_images: nil, use_main_content_only: nil, request_options: {})
+      # @overload web_scrape_md(url:, include_frames: nil, include_images: nil, include_links: nil, max_age_ms: nil, pdf: nil, shorten_base64_images: nil, timeout_ms: nil, use_main_content_only: nil, wait_for_ms: nil, request_options: {})
       #
       # @param url [String] Full URL to scrape into LLM usable Markdown (must include http:// or https:// pr
       #
@@ -241,12 +264,16 @@ module ContextDev
       #
       # @param max_age_ms [Integer] Return a cached result if a prior scrape for the same parameters exists and is y
       #
-      # @param parse_pdf [Boolean] When true (default), PDF URLs are fetched and their text layer is extracted and
+      # @param pdf [ContextDev::Models::WebWebScrapeMdParams::Pdf] PDF parsing controls. Use start/end to limit text extraction and OCR to an inclu
       #
       # @param shorten_base64_images [Boolean] Shorten base64-encoded image data in the Markdown output
       #
+      # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
+      #
       # @param use_main_content_only [Boolean] Extract only the main content of the page, excluding headers, footers, sidebars,
       #
+      # @param wait_for_ms [Integer] Optional browser wait time in milliseconds after initial page load before conver
+      #
       # @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}, nil]
       #
       # @return [ContextDev::Models::WebWebScrapeMdResponse]
@@ -263,9 +290,10 @@ module ContextDev
             include_images: "includeImages",
             include_links: "includeLinks",
             max_age_ms: "maxAgeMs",
-            parse_pdf: "parsePDF",
             shorten_base64_images: "shortenBase64Images",
-            use_main_content_only: "useMainContentOnly"
+            timeout_ms: "timeoutMS",
+            use_main_content_only: "useMainContentOnly",
+            wait_for_ms: "waitForMs"
           ),
           model: ContextDev::Models::WebWebScrapeMdResponse,
           options: options
@@ -277,12 +305,14 @@ module ContextDev
       #
       # Crawl an entire website's sitemap and return all discovered page URLs.
       #
-      # @overload web_scrape_sitemap(domain:, max_links: nil, url_regex: nil, request_options: {})
+      # @overload web_scrape_sitemap(domain:, max_links: nil, timeout_ms: nil, url_regex: nil, request_options: {})
       #
       # @param domain [String] Domain to build a sitemap for
       #
       # @param max_links [Integer] Maximum number of links to return from the sitemap crawl. Defaults to 10,000. Mi
       #
+      # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
+      #
       # @param url_regex [String] Optional RE2-compatible regex pattern. Only URLs matching this pattern are retur
       #
       # @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}, nil]
@@ -296,7 +326,7 @@ module ContextDev
         @client.request(
           method: :get,
           path: "web/scrape/sitemap",
-          query: query.transform_keys(max_links: "maxLinks", url_regex: "urlRegex"),
+          query: query.transform_keys(max_links: "maxLinks", timeout_ms: "timeoutMS", url_regex: "urlRegex"),
           model: ContextDev::Models::WebWebScrapeSitemapResponse,
           options: options
         )

data/lib/context_dev/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module ContextDev
-  VERSION = "1.16.0"
+  VERSION = "1.18.0"
 end

data/rbi/context_dev/models/ai_extract_product_params.rbi

@@ -27,8 +27,9 @@ module ContextDev
       sig { params(max_age_ms: Integer).void }
       attr_writer :max_age_ms
 
-      # Optional timeout in milliseconds for the request. Maximum allowed value is
-      # 300000ms (5 minutes).
+      # Optional timeout in milliseconds for the request. If the request takes longer
+      # than this value, it will be aborted with a 408 status code. Maximum allowed
+      # value is 300000ms (5 minutes).
       sig { returns(T.nilable(Integer)) }
       attr_reader :timeout_ms
 
@@ -50,8 +51,9 @@ module ContextDev
         # younger than this many milliseconds. Defaults to 7 days (604800000 ms) when
         # omitted. Max is 30 days (2592000000 ms). Set to 0 to always scrape fresh.
         max_age_ms: nil,
-        # Optional timeout in milliseconds for the request. Maximum allowed value is
-        # 300000ms (5 minutes).
+        # Optional timeout in milliseconds for the request. If the request takes longer
+        # than this value, it will be aborted with a 408 status code. Maximum allowed
+        # value is 300000ms (5 minutes).
         timeout_ms: nil,
         request_options: {}
       )

data/rbi/context_dev/models/ai_extract_products_params.rbi

@@ -92,8 +92,9 @@ module ContextDev
           sig { params(max_products: Integer).void }
           attr_writer :max_products
 
-          # Optional timeout in milliseconds for the request. Maximum allowed value is
-          # 300000ms (5 minutes).
+          # Optional timeout in milliseconds for the request. If the request takes longer
+          # than this value, it will be aborted with a 408 status code. Maximum allowed
+          # value is 300000ms (5 minutes).
           sig { returns(T.nilable(Integer)) }
           attr_reader :timeout_ms
 
@@ -117,8 +118,9 @@ module ContextDev
             max_age_ms: nil,
             # Maximum number of products to extract.
             max_products: nil,
-            # Optional timeout in milliseconds for the request. Maximum allowed value is
-            # 300000ms (5 minutes).
+            # Optional timeout in milliseconds for the request. If the request takes longer
+            # than this value, it will be aborted with a 408 status code. Maximum allowed
+            # value is 300000ms (5 minutes).
             timeout_ms: nil
           )
           end
@@ -167,8 +169,9 @@ module ContextDev
           sig { params(max_products: Integer).void }
           attr_writer :max_products
 
-          # Optional timeout in milliseconds for the request. Maximum allowed value is
-          # 300000ms (5 minutes).
+          # Optional timeout in milliseconds for the request. If the request takes longer
+          # than this value, it will be aborted with a 408 status code. Maximum allowed
+          # value is 300000ms (5 minutes).
           sig { returns(T.nilable(Integer)) }
           attr_reader :timeout_ms
 
@@ -193,8 +196,9 @@ module ContextDev
             max_age_ms: nil,
             # Maximum number of products to extract.
             max_products: nil,
-            # Optional timeout in milliseconds for the request. Maximum allowed value is
-            # 300000ms (5 minutes).
+            # Optional timeout in milliseconds for the request. If the request takes longer
+            # than this value, it will be aborted with a 408 status code. Maximum allowed
+            # value is 300000ms (5 minutes).
             timeout_ms: nil
           )
           end

data/rbi/context_dev/models/web_screenshot_params.rbi

@@ -69,22 +69,14 @@ module ContextDev
       sig { params(page: ContextDev::WebScreenshotParams::Page::OrSymbol).void }
       attr_writer :page
 
-      # Optional parameter to prioritize screenshot capture. If 'speed', optimizes for
-      # faster capture with basic quality. If 'quality', optimizes for higher quality
-      # with longer wait times. Defaults to 'quality' if not provided.
-      sig do
-        returns(
-          T.nilable(ContextDev::WebScreenshotParams::Prioritize::OrSymbol)
-        )
-      end
-      attr_reader :prioritize
+      # Optional timeout in milliseconds for the request. If the request takes longer
+      # than this value, it will be aborted with a 408 status code. Maximum allowed
+      # value is 300000ms (5 minutes).
+      sig { returns(T.nilable(Integer)) }
+      attr_reader :timeout_ms
 
-      sig do
-        params(
-          prioritize: ContextDev::WebScreenshotParams::Prioritize::OrSymbol
-        ).void
-      end
-      attr_writer :prioritize
+      sig { params(timeout_ms: Integer).void }
+      attr_writer :timeout_ms
 
       # Optional browser viewport dimensions for the screenshot. Defaults to 1920x1080.
       sig { returns(T.nilable(ContextDev::WebScreenshotParams::Viewport)) }
@@ -95,6 +87,15 @@ module ContextDev
       end
       attr_writer :viewport
 
+      # Optional browser wait time in milliseconds after initial page load before taking
+      # the screenshot. Min: 0. Max: 30000 (30 seconds). Defaults to 3000 ms when
+      # omitted.
+      sig { returns(T.nilable(Integer)) }
+      attr_reader :wait_for_ms
+
+      sig { params(wait_for_ms: Integer).void }
+      attr_writer :wait_for_ms
+
       sig do
         params(
           direct_url: String,
@@ -103,8 +104,9 @@ module ContextDev
             ContextDev::WebScreenshotParams::FullScreenshot::OrSymbol,
           max_age_ms: Integer,
           page: ContextDev::WebScreenshotParams::Page::OrSymbol,
-          prioritize: ContextDev::WebScreenshotParams::Prioritize::OrSymbol,
+          timeout_ms: Integer,
           viewport: ContextDev::WebScreenshotParams::Viewport::OrHash,
+          wait_for_ms: Integer,
           request_options: ContextDev::RequestOptions::OrHash
         ).returns(T.attached_class)
       end
@@ -131,12 +133,16 @@ module ContextDev
         # provided, screenshots the main domain landing page. Only applicable when using
         # 'domain', not 'directUrl'.
         page: nil,
-        # Optional parameter to prioritize screenshot capture. If 'speed', optimizes for
-        # faster capture with basic quality. If 'quality', optimizes for higher quality
-        # with longer wait times. Defaults to 'quality' if not provided.
-        prioritize: nil,
+        # Optional timeout in milliseconds for the request. If the request takes longer
+        # than this value, it will be aborted with a 408 status code. Maximum allowed
+        # value is 300000ms (5 minutes).
+        timeout_ms: nil,
         # Optional browser viewport dimensions for the screenshot. Defaults to 1920x1080.
         viewport: nil,
+        # Optional browser wait time in milliseconds after initial page load before taking
+        # the screenshot. Min: 0. Max: 30000 (30 seconds). Defaults to 3000 ms when
+        # omitted.
+        wait_for_ms: nil,
         request_options: {}
       )
       end
@@ -150,8 +156,9 @@ module ContextDev
               ContextDev::WebScreenshotParams::FullScreenshot::OrSymbol,
             max_age_ms: Integer,
             page: ContextDev::WebScreenshotParams::Page::OrSymbol,
-            prioritize: ContextDev::WebScreenshotParams::Prioritize::OrSymbol,
+            timeout_ms: Integer,
             viewport: ContextDev::WebScreenshotParams::Viewport,
+            wait_for_ms: Integer,
             request_options: ContextDev::RequestOptions
           }
         )
@@ -230,38 +237,6 @@ module ContextDev
         end
       end
 
-      # Optional parameter to prioritize screenshot capture. If 'speed', optimizes for
-      # faster capture with basic quality. If 'quality', optimizes for higher quality
-      # with longer wait times. Defaults to 'quality' if not provided.
-      module Prioritize
-        extend ContextDev::Internal::Type::Enum
-
-        TaggedSymbol =
-          T.type_alias do
-            T.all(Symbol, ContextDev::WebScreenshotParams::Prioritize)
-          end
-        OrSymbol = T.type_alias { T.any(Symbol, String) }
-
-        SPEED =
-          T.let(
-            :speed,
-            ContextDev::WebScreenshotParams::Prioritize::TaggedSymbol
-          )
-        QUALITY =
-          T.let(
-            :quality,
-            ContextDev::WebScreenshotParams::Prioritize::TaggedSymbol
-          )
-
-        sig do
-          override.returns(
-            T::Array[ContextDev::WebScreenshotParams::Prioritize::TaggedSymbol]
-          )
-        end
-        def self.values
-        end
-      end
-
       class Viewport < ContextDev::Internal::Type::BaseModel
         OrHash =
           T.type_alias do

data/rbi/context_dev/models/web_web_crawl_md_params.rbi

@@ -69,14 +69,13 @@ module ContextDev
       sig { params(max_pages: Integer).void }
       attr_writer :max_pages
 
-      # When true (default), PDF pages are fetched and their text layer is extracted and
-      # converted to Markdown alongside HTML pages. When false, PDF pages are skipped
-      # entirely (not included in results and not counted as failures).
-      sig { returns(T.nilable(T::Boolean)) }
-      attr_reader :parse_pdf
+      # PDF parsing controls. Use start/end to limit text extraction and OCR to an
+      # inclusive 1-based page range.
+      sig { returns(T.nilable(ContextDev::WebWebCrawlMdParams::Pdf)) }
+      attr_reader :pdf
 
-      sig { params(parse_pdf: T::Boolean).void }
-      attr_writer :parse_pdf
+      sig { params(pdf: ContextDev::WebWebCrawlMdParams::Pdf::OrHash).void }
+      attr_writer :pdf
 
       # Truncate base64-encoded image data in the Markdown output
       sig { returns(T.nilable(T::Boolean)) }
@@ -85,6 +84,25 @@ module ContextDev
       sig { params(shorten_base64_images: T::Boolean).void }
       attr_writer :shorten_base64_images
 
+      # Soft time budget for the crawl in milliseconds. After each scrape, the crawler
+      # checks the elapsed time and, if exceeded, returns the pages collected so far
+      # instead of continuing. Min: 10000 (10s). Max: 240000 (4 min). Default: 120000 (2
+      # min).
+      sig { returns(T.nilable(Integer)) }
+      attr_reader :stop_after_ms
+
+      sig { params(stop_after_ms: Integer).void }
+      attr_writer :stop_after_ms
+
+      # Optional timeout in milliseconds for the request. If the request takes longer
+      # than this value, it will be aborted with a 408 status code. Maximum allowed
+      # value is 300000ms (5 minutes).
+      sig { returns(T.nilable(Integer)) }
+      attr_reader :timeout_ms
+
+      sig { params(timeout_ms: Integer).void }
+      attr_writer :timeout_ms
+
       # Regex pattern. Only URLs matching this pattern will be followed and scraped.
       sig { returns(T.nilable(String)) }
       attr_reader :url_regex
@@ -100,6 +118,14 @@ module ContextDev
       sig { params(use_main_content_only: T::Boolean).void }
       attr_writer :use_main_content_only
 
+      # Optional browser wait time in milliseconds after initial page load for each
+      # crawled page. Min: 0. Max: 30000 (30 seconds).
+      sig { returns(T.nilable(Integer)) }
+      attr_reader :wait_for_ms
+
+      sig { params(wait_for_ms: Integer).void }
+      attr_writer :wait_for_ms
+
       sig do
         params(
           url: String,
@@ -110,10 +136,13 @@ module ContextDev
           max_age_ms: Integer,
           max_depth: Integer,
           max_pages: Integer,
-          parse_pdf: T::Boolean,
+          pdf: ContextDev::WebWebCrawlMdParams::Pdf::OrHash,
           shorten_base64_images: T::Boolean,
+          stop_after_ms: Integer,
+          timeout_ms: Integer,
           url_regex: String,
           use_main_content_only: T::Boolean,
+          wait_for_ms: Integer,
           request_options: ContextDev::RequestOptions::OrHash
         ).returns(T.attached_class)
       end
@@ -139,17 +168,28 @@ module ContextDev
         max_depth: nil,
         # Maximum number of pages to crawl. Hard cap: 500.
         max_pages: nil,
-        # When true (default), PDF pages are fetched and their text layer is extracted and
-        # converted to Markdown alongside HTML pages. When false, PDF pages are skipped
-        # entirely (not included in results and not counted as failures).
-        parse_pdf: nil,
+        # PDF parsing controls. Use start/end to limit text extraction and OCR to an
+        # inclusive 1-based page range.
+        pdf: nil,
         # Truncate base64-encoded image data in the Markdown output
         shorten_base64_images: nil,
+        # Soft time budget for the crawl in milliseconds. After each scrape, the crawler
+        # checks the elapsed time and, if exceeded, returns the pages collected so far
+        # instead of continuing. Min: 10000 (10s). Max: 240000 (4 min). Default: 120000 (2
+        # min).
+        stop_after_ms: nil,
+        # Optional timeout in milliseconds for the request. If the request takes longer
+        # than this value, it will be aborted with a 408 status code. Maximum allowed
+        # value is 300000ms (5 minutes).
+        timeout_ms: nil,
         # Regex pattern. Only URLs matching this pattern will be followed and scraped.
         url_regex: nil,
         # Extract only the main content, stripping headers, footers, sidebars, and
         # navigation
         use_main_content_only: nil,
+        # Optional browser wait time in milliseconds after initial page load for each
+        # crawled page. Min: 0. Max: 30000 (30 seconds).
+        wait_for_ms: nil,
         request_options: {}
       )
       end
@@ -165,16 +205,81 @@ module ContextDev
             max_age_ms: Integer,
             max_depth: Integer,
             max_pages: Integer,
-            parse_pdf: T::Boolean,
+            pdf: ContextDev::WebWebCrawlMdParams::Pdf,
             shorten_base64_images: T::Boolean,
+            stop_after_ms: Integer,
+            timeout_ms: Integer,
             url_regex: String,
             use_main_content_only: T::Boolean,
+            wait_for_ms: Integer,
             request_options: ContextDev::RequestOptions
           }
         )
       end
       def to_hash
       end
+
+      class Pdf < ContextDev::Internal::Type::BaseModel
+        OrHash =
+          T.type_alias do
+            T.any(
+              ContextDev::WebWebCrawlMdParams::Pdf,
+              ContextDev::Internal::AnyHash
+            )
+          end
+
+        # Last 1-based PDF page to parse. When omitted, parsing ends at the last page.
+        # Must be greater than or equal to start when both are provided.
+        sig { returns(T.nilable(Integer)) }
+        attr_reader :end_
+
+        sig { params(end_: Integer).void }
+        attr_writer :end_
+
+        # When true, PDF pages are fetched and parsed. When false, PDF pages are skipped
+        # entirely (not included in results and not counted as failures).
+        sig { returns(T.nilable(T::Boolean)) }
+        attr_reader :should_parse
+
+        sig { params(should_parse: T::Boolean).void }
+        attr_writer :should_parse
+
+        # First 1-based PDF page to parse. When omitted, parsing starts at the first page.
+        sig { returns(T.nilable(Integer)) }
+        attr_reader :start
+
+        sig { params(start: Integer).void }
+        attr_writer :start
+
+        # PDF parsing controls. Use start/end to limit text extraction and OCR to an
+        # inclusive 1-based page range.
+        sig do
+          params(
+            end_: Integer,
+            should_parse: T::Boolean,
+            start: Integer
+          ).returns(T.attached_class)
+        end
+        def self.new(
+          # Last 1-based PDF page to parse. When omitted, parsing ends at the last page.
+          # Must be greater than or equal to start when both are provided.
+          end_: nil,
+          # When true, PDF pages are fetched and parsed. When false, PDF pages are skipped
+          # entirely (not included in results and not counted as failures).
+          should_parse: nil,
+          # First 1-based PDF page to parse. When omitted, parsing starts at the first page.
+          start: nil
+        )
+        end
+
+        sig do
+          override.returns(
+            { end_: Integer, should_parse: T::Boolean, start: Integer }
+          )
+        end
+        def to_hash
+        end
+      end
     end
   end
 end

data/rbi/context_dev/models/web_web_crawl_md_response.rbi

@@ -64,7 +64,8 @@ module ContextDev
         sig { returns(Integer) }
         attr_accessor :num_failed
 
-        # Number of URLs skipped (PDFs when parsePDF=false, or URLs not matching urlRegex)
+        # Number of URLs skipped (PDFs when pdf.shouldParse=false, or URLs not matching
+        # urlRegex)
         sig { returns(Integer) }
         attr_accessor :num_skipped
 
@@ -90,7 +91,8 @@ module ContextDev
           max_crawl_depth:,
           # Number of pages that failed to crawl
           num_failed:,
-          # Number of URLs skipped (PDFs when parsePDF=false, or URLs not matching urlRegex)
+          # Number of URLs skipped (PDFs when pdf.shouldParse=false, or URLs not matching
+          # urlRegex)
           num_skipped:,
           # Number of pages successfully crawled
           num_succeeded:,