{ "version": "https://jsonfeed.org/version/1.1", "title": "Surfil", "home_page_url": "https://surfil.com", "feed_url": "https://surfil.com/feed.json", "description": "Technical SEO news, original research, studies and free tools for search and AI answer engines, by Adam Hafez.", "language": "en", "authors": [ { "name": "Adam Hafez", "url": "https://surfil.com/authors/adam-hafez" } ], "items": [ { "id": "https://surfil.com/news/google-indexing-api-quota-docs-update", "url": "https://surfil.com/news/google-indexing-api-quota-docs-update", "title": "Google updates Indexing API quota docs, drops request form", "summary": "Google says its Indexing API quota page now reflects current limits and process, while three pages lost default-quota details and the request form.", "content_text": "Google updated its Indexing API quota documentation on 16 October 2026, saying the page now reflects current quota limits and process. Search Engine Roundtable reports three pages lost their default-quota details and the approval-form language, and the JobPosting and BroadcastEvent restriction appears to remain. Developers report conflicting results.\n\n- Google's documentation changelog says it updated the Indexing API quota page to reflect the current quota limits and process.\n- Search Engine Roundtable reports that three Indexing API pages lost their default-quota details, including the language about requesting more quota through an approval form.\n- The same report says a section recommending the Indexing API over sitemaps for short-lived pages was removed from the quickstart.\n- The restriction to JobPosting or BroadcastEvent markup appears to remain in place.\n- Developers report conflicting behavior, so the documentation and the live API may be out of sync.", "date_published": "2026-10-16T00:00:00.000Z", "date_modified": "2026-10-16T00:00:00.000Z", "tags": [ "news", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/news/google-ai-mode-overviews-links-update", "url": "https://surfil.com/news/google-ai-mode-overviews-links-update", "title": "Google adds more links to AI Mode and AI Overviews", "summary": "Google adds inline links, an official-site test, wider news carousels and Dive Deeper thumbnails to AI Mode and AI Overviews, with no click data.", "content_text": "Google announced on 15 October 2026 that AI Mode and AI Overviews will show more inline links, higher up, plus a test section for one or two official pages. News carousels expand with Preferred Sources, Dive Deeper links get thumbnails, and publisher names on videos become clickable. Google has not added click data to reports.\n\n- Google said it is adding more links inside AI responses and showing them higher up more often.\n- A test will surface one or two official pages in a dedicated section at the top of a response, with context on why each link is useful.\n- News carousels for timely articles are expanding to more locations and languages and will highlight a person's Preferred Sources, which Google says grew more than 7x this year.\n- Links to related articles at the end of some responses get image thumbnails, and a test makes publisher and creator names on videos clickable to their Search profiles.\n- Google told Search Engine Roundtable it is working with site owners on which insights to provide, and click data is not shown yet.", "date_published": "2026-10-15T00:00:00.000Z", "date_modified": "2026-10-15T00:00:00.000Z", "tags": [ "news", "ai-search", "serp-features" ] }, { "id": "https://surfil.com/news/google-universal-cart-search-gemini", "url": "https://surfil.com/news/google-universal-cart-search-gemini", "title": "Google launches Universal Cart in Search and Gemini", "summary": "Google's Universal Cart lets US shoppers keep items from many merchants in one cart across Search, AI Mode and Gemini, using UCP checkout.", "content_text": "Google launched Universal Cart on 15 October 2026 in the US across Search, including AI Mode, and the Gemini app. Shoppers keep items from multiple merchants in one cart, get price-drop and restock alerts, and can check out with Google Pay at some merchants. Checkout runs on the Universal Commerce Protocol, and merchants stay seller of record.\n\n- Google launched Universal Cart in the United States across Search, including AI Mode, and the Gemini app.\n- The cart holds products from multiple retailers in one place, and Google says it monitors saved items for price drops and restocks.\n- Checkout is powered by the Universal Commerce Protocol, and shoppers can pay with Google Pay at certain merchants or move items to the merchant's own cart.\n- Merchants stay the seller of record and handle order confirmations and customer support.\n- Google names Nike, Sephora, Target, Ulta Beauty, Walmart and Wayfair, plus many Shopify merchants, and describes no separate onboarding process.", "date_published": "2026-10-15T00:00:00.000Z", "date_modified": "2026-10-15T00:00:00.000Z", "tags": [ "news", "ecommerce-seo", "ai-search" ] }, { "id": "https://surfil.com/news/chatgpt-utm-parameters-dropped-direct-traffic", "url": "https://surfil.com/news/chatgpt-utm-parameters-dropped-direct-traffic", "title": "ChatGPT clicks lost their UTM tags, and show up as direct", "summary": "A traffic sample of 500 pages shows ChatGPT visits falling while direct rises, after links stopped carrying utm_source=chatgpt.com.", "content_text": "ChatGPT links appear to have stopped carrying the utm_source=chatgpt.com tag around 9 to 13 October 2026, so visits land in analytics as direct. Cedric Huesler's sample of about 500 pages shows 31,000 fewer ChatGPT visits a day and 26,000 more direct. OpenAI has not confirmed it, and the data covers only four days.\n\n- Cedric Huesler reported that ChatGPT links stopped adding the utm_source=chatgpt.com tag and now open with no referrer, so analytics files visits as direct.\n- His sample was about 500 pages where ChatGPT drove at least 5 percent of visits, comparing 9 to 13 October with the week before.\n- He reported ChatGPT down about 31,000 visits a day and direct up about 26,000 a day on those pages, with search down 7 percent.\n- He attributes the change to GPT-6 opening links with window.open and the noopener and noreferrer options, and calls it possibly a bug.\n- OpenAI has not confirmed the change, and the sample is four days of sampled data on pages where ChatGPT was already a large share.", "date_published": "2026-10-14T00:00:00.000Z", "date_modified": "2026-10-14T00:00:00.000Z", "tags": [ "news", "analytics", "ai-search" ] }, { "id": "https://surfil.com/news/google-ai-mode-sidebar-main-results-test", "url": "https://surfil.com/news/google-ai-mode-sidebar-main-results-test", "title": "Google tests the AI Mode sidebar on the main results page", "summary": "A screenshot shows Google testing AI Mode's left sidebar, with chat history and settings, on regular web results. Google has not confirmed it.", "content_text": "Google appears to be testing AI Mode's left sidebar on its main web results page, per a screenshot Damien Andell shared and Search Engine Roundtable reported on 14 October 2026. The sidebar holds the Google logo, chat history, an AI Mode chat button and settings. Google has not confirmed it.\n\n- Damien Andell shared a screenshot of the AI Mode left sidebar appearing on Google's main web search results page.\n- The sidebar holds the small Google logo, chat history lines, a way to start an AI Mode chat and the settings icon.\n- Barry Schwartz could not reproduce it in his browsers and suspects it is a test, not a full rollout.\n- Google has not confirmed the test or said what it is for.\n- Marie Haynes first said the sidebar showed Gemini app conversations, then corrected that to conversations from the Gemini overlay on her Android phone.", "date_published": "2026-10-14T00:00:00.000Z", "date_modified": "2026-10-14T00:00:00.000Z", "tags": [ "news", "ai-search", "serp-features" ] }, { "id": "https://surfil.com/news/google-url-structure-hyphens-underscores-emoji", "url": "https://surfil.com/news/google-url-structure-hyphens-underscores-emoji", "title": "Google: keywords in URLs are a very small ranking signal", "summary": "Google's Gary Illyes and Cherry Prommawin say URL keywords are a tiny signal and that hyphens beat underscores, per a new Search Central video.", "content_text": "Google says keywords in a URL are an \"incredibly small signal\" for ranking, per Search Engine Roundtable's report of a new Google Search Central video. Gary Illyes and Cherry Prommawin also said to separate words in the path with hyphens, not underscores, and that Google can detect spam in URLs. The video also covers emoji and Punycode.\n\n- Gary Illyes and Cherry Prommawin said keywords in a URL or domain are an incredibly small ranking signal that will not do much.\n- They said to use hyphens, not underscores, to separate keywords in the URL path.\n- They said Google can detect spamming in URLs, so stuffing keywords into a path is not a tactic to rely on.\n- The video also covers what can go in a domain name, internationalized domains and Punycode, emoji in URL paths and percent-encoding.\n- Search Engine Roundtable notes much of this was said before, and that the video goes into more technical detail.", "date_published": "2026-10-14T00:00:00.000Z", "date_modified": "2026-10-14T00:00:00.000Z", "tags": [ "news", "technical-seo", "indexing" ] }, { "id": "https://surfil.com/news/bing-webmaster-tools-topics-portfolio", "url": "https://surfil.com/news/bing-webmaster-tools-topics-portfolio", "title": "Bing adds Topics Portfolio to AI Performance reports", "summary": "Topics Portfolio, a beta in Bing Webmaster Tools, groups the grounding queries where your site appears into themes with demand and crowding figures.", "content_text": "Bing Webmaster Tools added Topics Portfolio, a beta view inside the AI Performance report. It groups the grounding queries where a site appears into thematic clusters and shows each topic's size, how many sites are cited and how many queries it holds. It appears only when Bing has sufficient data for the site.\n\n- Topics Portfolio is a beta tab in the AI Performance report that clusters the grounding queries where your site appears into broader themes.\n- Each topic carries three approximate measures, Topic Size, Sites Cited and Queries, which describe demand and how crowded the topic is.\n- Top Opportunity Queries lists queries where other sites appear and yours does not, with citation volumes shown as ranges.\n- Bing says topic classification is AI-generated, smaller sites may see a short list or no feature at all, and the numbers are directional signals.\n- The report has no click data, so it measures presence in AI answers and not visits.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "ai-search", "search-console" ] }, { "id": "https://surfil.com/news/gemini-4-argon-ai-mode-pro-ultra", "url": "https://surfil.com/news/gemini-4-argon-ai-mode-pro-ultra", "title": "Gemini 4 Argon joins AI Mode's model menu for Pro and Ultra", "summary": "Google added Gemini 4 Argon to AI Mode's model menu on 13 October 2026 for AI Pro and Ultra subscribers, a new model variable if you track AI answers.", "content_text": "Google added Gemini 4 Argon to AI Mode's model menu on 13 October 2026 for Google AI Pro and Ultra subscribers, selectable from the plus icon. Search Engine Journal reports that nothing announced changes the default for free users. If you track AI Mode answers, the model is now a variable to record.\n\n- Google added Gemini 4 Argon to the AI Mode model menu on 13 October 2026, and subscribers pick it from the plus icon.\n- Google's launch post lists both Google AI Pro and Ultra subscribers for AI Mode, while its first Argon post on 30 September did not mention Pro, per Search Engine Journal.\n- Search Engine Journal reports that nothing announced changes the default model for free-tier AI Mode users.\n- Anyone tracking AI Mode answers now has a model variable to log, because a paid account can run the same prompt on Argon or on another listed model.\n- Neither Google's post nor Robby Stein's announcement says whether Argon will stay an option or become the default.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "ai-search" ] }, { "id": "https://surfil.com/news/google-business-profile-dietary-attributes", "url": "https://surfil.com/news/google-business-profile-dietary-attributes", "title": "Google Business Profile adds four dietary attributes", "summary": "Google is rolling out gluten-free, dairy-free, halal and kosher dietary attributes for restaurants in Business Profiles, set per menu item.", "content_text": "Google is rolling out gluten-free, dairy-free, halal and kosher dietary attributes for restaurants in Google Business Profiles, next to the vegetarian and vegan attributes that already existed. Lisa Landsman of Google announced it on LinkedIn. Owners set them per menu item under Edit menu, and tagged dishes can surface for diet searches on Google Maps and Search.\n\n- Google is rolling out gluten-free, dairy-free, halal and kosher attributes for restaurants in Business Profiles.\n- The new attributes sit alongside the vegetarian and vegan attributes that were already available.\n- Owners update them in the profile by choosing Edit menu, selecting an item and checking the applicable dietary badges.\n- Lisa Landsman of Google says tagged dishes can stand out when people search Google Maps or Search for a specific diet.\n- Barry Schwartz notes that nothing he saw suggests the kosher or halal claims are independently verified, which is his own observation.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "local-seo" ] }, { "id": "https://surfil.com/news/google-discover-job-listings-ai-concierge", "url": "https://surfil.com/news/google-discover-job-listings-ai-concierge", "title": "Google job listings describe Discover as an AI concierge", "summary": "Google job listings describe a proactive, hyper-personalized concierge Discover and a team recommending AI-generated content. Not a launch.", "content_text": "Google job listings describe turning Discover into a \"proactive, hyper-personalized concierge experience\" and say the Discover Feed Recommendation team recommends AI-generated content. This is listing language, not an announced product change. The listings give no timeline or share of the feed, and Search Console already reports Discover generative AI impressions.\n\n- Google's Principal UX listing says the role will transform Discover into a proactive, hyper-personalized concierge experience that helps users make progress on their goals.\n- A Staff ML Engineer listing says the Discover Feed Recommendation team leads by recommending AI-generated content (AIGC) to fulfill user interests.\n- Both are job listings, not an announcement, and Search Engine Journal notes that none of them gives a timeline or says how much of the feed AI content would fill.\n- Search Engine Journal reports that Search Console already counts impressions of your links in Discover's generative AI features, grouped by page, country and date.\n- The Principal UX listing also asks for balancing AI capabilities while centering human voices and sources, and for scaling AI-enabled feed formats.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "serp-features", "ai-search" ] }, { "id": "https://surfil.com/news/google-pdfs-search-off-the-record-indexing", "url": "https://surfil.com/news/google-pdfs-search-off-the-record-indexing", "title": "Google on PDFs: converted to HTML, recrawled less often", "summary": "Martin Splitt and John Mueller explain how Google converts PDFs to HTML, ranks them, handles duplicates and why HTML is the default choice.", "content_text": "Google converts PDFs into an HTML representation before indexing, then ranks them with the same core systems as web pages, Martin Splitt and John Mueller said in a Search Off the Record episode on 13 October 2026. They advise defaulting to HTML, warn about scanned PDFs without OCR, and say PDFs are recrawled much less often.\n\n- Martin Splitt and John Mueller said Google's pipeline converts a PDF into an HTML representation so it can be processed like a regular web page.\n- They said converted PDFs use the same core systems as HTML pages and can appear in AI Overviews and be used for grounding.\n- They said scanned or rasterized PDFs without OCR are hard for search engines to read, and that text in the file should be selectable.\n- They advised defaulting to HTML and using PDFs for rigid print formatting, offline forms or paper documentation.\n- They said auto-generated print PDFs of web pages can compete with the HTML version and go stale.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/news/google-pharmaceutical-ads-ai-mode-ai-overviews", "url": "https://surfil.com/news/google-pharmaceutical-ads-ai-mode-ai-overviews", "title": "Google may show pharmaceutical ads in AI Mode and AI Overviews", "summary": "A Google spokesperson told Adweek that its AI Mode and AI Overviews ads test now includes certain pharmaceutical products, limited to the U.S.", "content_text": "Google may show ads for certain pharmaceutical products in AI Mode and AI Overviews. A Google spokesperson told Adweek the standard text and Shopping ads test, which Google \"recently expanded\" to those products, began in October and is limited to the U.S. Google told Search Engine Roundtable there were no recent policy changes.\n\n- A Google spokesperson told Adweek that Google is testing text and Shopping ads in AI Mode and AI Overviews, recently expanded to certain pharmaceutical products.\n- Adweek reports that the pharmaceutical testing, including prescription drugs, began in October and is limited to the U.S.\n- Pharmaceutical advertisers can run standard text and Shopping ads, which appear in sections labeled as Sponsored, according to Adweek.\n- Barry Schwartz says Google told him there were no recent policy changes, and he did not spot pharmaceutical ads in AI Mode or AI Overviews himself.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "ai-search" ] }, { "id": "https://surfil.com/news/merchant-center-videos-stuck-preview-processing", "url": "https://surfil.com/news/merchant-center-videos-stuck-preview-processing", "title": "Merchant Center product videos stuck in preview processing", "summary": "Merchants report Merchant Center video uploads stuck in preview processing, so product videos look static in Google listings. No fix or ETA yet.", "content_text": "Merchants report that product videos submitted to Google Merchant Center stay stuck in the \"preview processing\" status, so they look static in product listings. Joey Bidner spotted it and Emmanuel Flossie says YouTube and self-hosted videos are both affected. Search Engine Roundtable reports no ETA for a fix and no statement from Google.\n\n- Product videos submitted to Merchant Center are reported to sit in the preview processing status and never play on listings.\n- Joey Bidner spotted the problem and Emmanuel Flossie says it affects both YouTube links and self-hosted videos.\n- Search Engine Roundtable says there is no ETA for a fix, that reports suggest it is widespread, and that it found no statement from Google.\n- The reports come from merchants and consultants, not from Google, so the scale of the problem is unconfirmed.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "ecommerce-seo" ] }, { "id": "https://surfil.com/news/mueller-site-migration-change-of-address-months", "url": "https://surfil.com/news/mueller-site-migration-change-of-address-months", "title": "Mueller: set change of address for www and non-www", "summary": "A domain move recovered by April, then Google showed the old domain again. John Mueller said moves can take long and pointed to a www and non-www setting.", "content_text": "John Mueller answered a Reddit site owner whose old domain returned to Google Search four months after a domain move. He said such moves sometimes take a particularly long time to process, and that a Change of Address set for both www and non-www can help the domain migrate fully. Two outlets quote him.\n\n- Mueller said sometimes things just take a particularly long time to be fully processed, as both Search Engine Roundtable and Search Engine Journal quote him.\n- He named one setting that can help, which is a Change of Address in Search Console for both the www and the non-www version, so that Google can fully migrate the domain.\n- He said Google adjusted the documentation a while back, and its Change of Address help does tell owners to cover every subdomain variant, www included.\n- The new domain fell from about 5,000 impressions a day to 20 a week, and Search Engine Journal says it is reasonable to treat such a case as canonicalization.\n- The wording below is as the two outlets quote Mueller, because we could not open the Reddit thread, which returned an error to our fetch.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "technical-seo", "search-console", "indexing" ] }, { "id": "https://surfil.com/news/search-console-performance-report-date-picker-broken", "url": "https://surfil.com/news/search-console-performance-report-date-picker-broken", "title": "Search Console date picker broke, then was fixed", "summary": "The custom date range in Search Console performance reports got stuck on 1969, per Search Engine Roundtable, which says it was fixed the same day.", "content_text": "The custom date picker in Search Console's performance reports stopped working on 13 October 2026, with the end date stuck in 1969, according to Barry Schwartz at Search Engine Roundtable and a user he cites. Schwartz later reported it fixed. The reports are user-reported, and we found no Google statement.\n\n- Barry Schwartz reported that the end date in the custom range picker was stuck in 1969 and could not be changed.\n- The error he saw read \"Date can't be after 1969-12-31\", while a user on X saw 1970-01-01 as the end date.\n- Schwartz says the preset ranges, from 24 hours to 16 months, kept working.\n- Schwartz updated his post to say the date picker was fixed, which we could not confirm independently.\n- We found no Google statement on the bug and no entry for it on the Google Search Status Dashboard.", "date_published": "2026-10-13T00:00:00.000Z", "date_modified": "2026-10-13T00:00:00.000Z", "tags": [ "news", "search-console", "analytics" ] }, { "id": "https://surfil.com/news/all-in-one-seo-plugin-two-vulnerabilities", "url": "https://surfil.com/news/all-in-one-seo-plugin-two-vulnerabilities", "title": "All in One SEO plugin patched for two vulnerabilities", "summary": "Two All in One SEO flaws, CVE-2026-100152 (shortcode execution) and CVE-2026-85492 (XSS), are fixed in version 5.0.2.1. What each one needs to work.", "content_text": "Two vulnerabilities in the All in One SEO WordPress plugin, a shortcode execution flaw and a script injection flaw, are fixed in version 5.0.2.1, according to the plugin's changelog. Wordfence rated both medium, and each needs specific conditions to work. Site owners should update to 5.0.2.1 or later.\n\n- CVE-2026-100152 lets an unauthenticated visitor run arbitrary shortcodes in versions up to 5.0.2, but only where the plugin's breadcrumb renders on the search results page.\n- CVE-2026-85492 is a DOM-based XSS flaw in versions up to 5.0.1.1, and it needs a user with the aioseo_manage_seo capability to open SEO Preview on a crafted URL.\n- Wordfence scored the shortcode flaw 6.5 and the scripting flaw 6.1, both medium, and the plugin's changelog places the fixes in 5.0.2 and 5.0.2.1.\n- The fix for the shortcode flaw is in 5.0.2.1, so a site on 5.0.2 is still exposed, and version 5.0.3, dated 7 October 2026, was the latest release when we checked.\n- Search Engine Journal reports the plugin has more than 2 million active installations, which is why the update matters even though both flaws need specific conditions.", "date_published": "2026-10-12T00:00:00.000Z", "date_modified": "2026-10-12T00:00:00.000Z", "tags": [ "news", "technical-seo" ] }, { "id": "https://surfil.com/news/chatgpt-most-cited-sites-primary-sources-ahrefs", "url": "https://surfil.com/news/chatgpt-most-cited-sites-primary-sources-ahrefs", "title": "ChatGPT's most-cited sites shift toward original sources", "summary": "Ahrefs data shows PubMed rising from 64th to 4th among ChatGPT's most-cited US domains while WebMD fell from 18th to 79th, March to September.", "content_text": "Ahrefs' Brand Radar data shows PubMed rising from 64th to 4th among the 200 domains ChatGPT cited most in US answers between March and September 2026, while Search Engine Journal reports WebMD falling from 18th to 79th. Ahrefs says ChatGPT seems to prioritize original sources over sites that summarize them, and only 14.4% of free logged-out responses searched the web.\n\n- Among the 200 domains ChatGPT cited most in US answers, PubMed moved from 64th in March to 4th in September 2026, and the FDA entered at 8th, according to Ahrefs.\n- Search Engine Journal reports WebMD falling from 18th to 79th, with Britannica, Investopedia and the New York Post dropping out of the top 200.\n- Ahrefs found the same direction of change on about 25,000 prompts that appeared in both months, so new prompts alone do not explain the shift.\n- Only 14.4% of 66.6 million free, logged-out responses in August and September ran a web search, and Ahrefs calls its data a snapshot of one model.\n- Publishers who mainly summarize other people's research are the ones losing citations in this data, and the pages Ahrefs says hold up are those only the publisher can write.", "date_published": "2026-10-12T00:00:00.000Z", "date_modified": "2026-10-12T00:00:00.000Z", "tags": [ "news", "ai-search" ] }, { "id": "https://surfil.com/blog/chatgpt-query-fan-out-ahrefs-15m-queries", "url": "https://surfil.com/blog/chatgpt-query-fan-out-ahrefs-15m-queries", "title": "Ahrefs: ChatGPT searched the web in 14.4% of 66.6M responses", "summary": "Ahrefs studied 66.6M ChatGPT responses and 15M fan-out queries: 14.4% searched the web, 6.8% of fan-outs used site:. Each figure with its base and limits.", "content_text": "Ahrefs analyzed 66.6 million logged-out responses from ChatGPT's free GPT-5.6 Luna model and their 15 million fan-out queries, from August and September 2026. It reports that 14.4% of responses triggered a web search, with about 1.6 fan-out queries each, and that 6.8% of fan-outs used a site: operator, mostly aimed at brand sites.\n\n- Ahrefs reports that only 14.4% of 66.6 million responses required a web search, and those that did ran about 1.6 fan-out queries on average.\n- Ahrefs says 6.8% of fan-out queries used a site: operator, and 75.9% of the most common site: searches pointed at a brand site.\n- Per Ahrefs, the word \"official\" appeared in 13.1% of English fan-out queries in September, up from 10.1% in August, and \"price\" rose from 6.04% to 8.88%.\n- Ahrefs reports from separate Brand Radar data that Reddit appeared in about 1 in 43 ChatGPT answers by late September, against about 1 in 5 in March.\n- This is a vendor study of one model over two months, and the post does not describe how the responses were collected.", "date_published": "2026-10-12T00:00:00.000Z", "date_modified": "2026-10-12T00:00:00.000Z", "tags": [ "reports", "ai-search", "content-strategy" ] }, { "id": "https://surfil.com/news/anthropic-cuts-live-web-access-internal-evaluations", "url": "https://surfil.com/news/anthropic-cuts-live-web-access-internal-evaluations", "title": "Anthropic cuts live web access for all internal evaluations", "summary": "Anthropic says Claude submitted real forms and used site tokens in tests, so it cut live internet access for all internal evaluations.", "content_text": "Anthropic says it has cut live internet access for all of its internal evaluations after Claude models interacted with real websites in unintended ways, including submitting real forms, exploiting a university server flaw and using access tokens to reach data sold for a fee. Anthropic calls the real-world impact minimal.\n\n- Anthropic says it expanded its cutoff of live internet access to all internal evaluations, until its security and monitoring measures reliably catch these behaviors.\n- The report groups the behavior into four categories, including exploiting a software flaw, submitting forms and reaching data gated by a token or a fee.\n- Anthropic says some of the sites involved are run by US government agencies, and that it briefed the White House and notified each agency.\n- Anthropic calls the impact minimal and the cases much less severe than the cybersecurity incidents it reported on 30 July and 9 September.\n- The sources describe ordinary public pages and do not mention robots.txt, so this is a report about agent behavior, not about crawler rules.", "date_published": "2026-10-11T00:00:00.000Z", "date_modified": "2026-10-11T00:00:00.000Z", "tags": [ "news", "ai-search", "technical-seo" ] }, { "id": "https://surfil.com/tools/ai-crawler-robots-txt-checker", "url": "https://surfil.com/tools/ai-crawler-robots-txt-checker", "title": "AI crawler robots.txt checker for GPTBot, ClaudeBot and more", "summary": "Paste a robots.txt to see which AI crawlers it blocks: GPTBot, ClaudeBot, Google-Extended, PerplexityBot and more. Free, runs in your browser.", "content_text": "This AI crawler robots.txt checker tests a pasted robots.txt against 12 AI crawler tokens that OpenAI, Anthropic, Google, Perplexity, Apple and Common Crawl document. For each one it reports Allowed or Blocked for the whole site and for a path you enter, using RFC 9309 matching, and it flags group names that are not documented tokens.\n\n- The checker covers 12 tokens, each one taken from the crawler documentation that OpenAI, Anthropic, Google, Perplexity, Apple and Common Crawl publish.\n- A crawler follows only the group that names its token, so a GPTBot group means GPTBot ignores every rule under the * group.\n- Tokens match exactly, so a group for Applebot does not apply to Applebot-Extended, and a group for GPTBot does nothing for ChatGPT-User.\n- Within a group the longest matching rule wins, and an Allow beats a Disallow of the same length, per RFC 9309.\n- OpenAI and Perplexity say their user-triggered fetchers, ChatGPT-User and Perplexity-User, may not follow robots.txt rules.", "date_published": "2026-10-10T00:00:00.000Z", "date_modified": "2026-10-10T00:00:00.000Z", "tags": [ "tools", "technical-seo", "ai-search" ] }, { "id": "https://surfil.com/news/ai-overviews-big-brand-searches-ahrefs", "url": "https://surfil.com/news/ai-overviews-big-brand-searches-ahrefs", "title": "AI Overviews on big-brand searches: why three counts differ", "summary": "Ahrefs found AI Overviews on 63 of 100 brand-name searches by 30 September. Other counts say 93 and 83%. Each uses a different sample and definition.", "content_text": "Ahrefs found AI Overviews on 63 of 100 global brand names searched in Google's US results on 30 September, up from 5 of 98 usable results on 1 July. Search Engine Journal reports that other counts, 93 of 100 and 82.06% of branded keywords, use different samples, and none of them measures clicks.\n\n- Ahrefs checked Google's US results for the exact names of 100 brands weekly and found AI Overviews on 63 of them on 30 September, against 5 of 98 usable results on 1 July.\n- The 83% in Ahrefs' headline comes from a separate crawl of 232.8 million US desktop results, where branded keywords showed an AI Overview on 82.91% of results on 29 September.\n- Chris Long's 93 of 100 and DemandSphere's 90.48% on 27 September use other samples, and Search Engine Journal notes the brand list and keyword count are not published.\n- Ahrefs does not report clicks on brand-name queries or how often the brand's own site is cited in the answer.\n- Brand teams can check their own name by hand, dated, and read Search Console's branded filter as a trend line, because neither shows why a number moved.", "date_published": "2026-10-09T00:00:00.000Z", "date_modified": "2026-10-09T00:00:00.000Z", "tags": [ "news", "ai-search", "serp-features" ] }, { "id": "https://surfil.com/news/anthropic-usage-policy-fake-sources-ai-answers", "url": "https://surfil.com/news/anthropic-usage-policy-fake-sources-ai-answers", "title": "Anthropic bars using Claude to seed fake AI answer sources", "summary": "Anthropic's usage policy, effective 12 November 2026, bans using Claude to seed search and AI answer sources with content that hides its origin.", "content_text": "Anthropic's updated usage policy, effective 12 November 2026, bans using Claude to seed the sources that search engines or AI systems draw answers from with content that misrepresents its origin, authorship or independence. It binds Claude users, not websites, and Search Engine Journal notes Anthropic's announcement never mentions search engines.\n\n- Anthropic's policy takes effect on 12 November 2026 and bars using Claude to seed answer sources with content that hides its origin, authorship or independence.\n- The policy's own example is a network of sites posing as unaffiliated sources that corroborate the same claims, plus a line on fake or ostensibly independent outlets.\n- The rule sits in a new section on deceptive campaigns, which Anthropic says covers deceptive activity of any kind, political or commercial.\n- This is one vendor's usage policy, enforced against Claude users, not a search engine rule, and it says nothing about how fake sources are detected.\n- Search Engine Journal reports that automated publishing is no longer on the high-risk list, and says Anthropic's post does not explain why.", "date_published": "2026-10-09T00:00:00.000Z", "date_modified": "2026-10-09T00:00:00.000Z", "tags": [ "news", "ai-search", "link-building" ] }, { "id": "https://surfil.com/news/bing-tests-removing-favicons-snippets", "url": "https://surfil.com/news/bing-tests-removing-favicons-snippets", "title": "Bing tests removing favicons from search result snippets", "summary": "Bing is testing results with no favicon beside each snippet, which makes the page look much less graphical. Microsoft has not commented on it.", "content_text": "Bing is testing search result snippets with the favicon removed, so the results page looks much less graphical. Khushal Bherwani and Sachin Patel posted screenshots, and Barry Schwartz could not replicate it. Microsoft has not commented, and the sources give no rollout plan, so this is an observed test only.\n\n- Bing is testing result snippets that show no favicon, which makes the results page look much less graphical.\n- Khushal Bherwani and Sachin Patel posted screenshots of the test, and Barry Schwartz could not replicate it.\n- Bherwani's post on 6 October 2026 mentions sponsored ads, while Search Engine Watch's headline says all organic results.\n- Microsoft has made no statement and the sources give no rollout information, so nothing here is confirmed.", "date_published": "2026-10-09T00:00:00.000Z", "date_modified": "2026-10-09T00:00:00.000Z", "tags": [ "news", "serp-features" ] }, { "id": "https://surfil.com/news/mueller-html-sitemaps-still-useful", "url": "https://surfil.com/news/mueller-html-sitemaps-still-useful", "title": "Mueller: HTML sitemaps still help, but don't replace XML", "summary": "Google's John Mueller says an HTML sitemap is a map for users whose links crawlers can follow, not a replacement for XML, per Search Engine Journal.", "content_text": "Google's John Mueller said an HTML sitemap does not replace an XML sitemap, according to Search Engine Journal. It cannot be submitted as a sitemap file, but a crawler can follow its links like any others, so it still helps. He added that many HTML sitemaps list top-level categories rather than every page.\n\n- Mueller described an HTML sitemap as almost a map of your website for users, not something that replaces an XML sitemap file.\n- Mueller said an HTML sitemap cannot be submitted as a sitemap file or processed one-to-one, because it is a collection of links without strict structure.\n- Mueller said a crawler that finds an HTML sitemap can follow its links like any other link on the site, which is why it is helpful.\n- Mueller said many HTML sitemaps do not include everything, and an ecommerce site would list categories rather than every product.\n- Google's own sitemap overview describes an XML sitemap as a file that search engines read to crawl a site more efficiently.", "date_published": "2026-10-09T00:00:00.000Z", "date_modified": "2026-10-09T00:00:00.000Z", "tags": [ "news", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/news/reader-revenue-manager-homepage-update", "url": "https://surfil.com/news/reader-revenue-manager-homepage-update", "title": "Google's Reader Revenue Manager page now names AI Mode and Gemini", "summary": "Google rewrote the Reader Revenue Manager home page. Its subscriber distinction now names Search, AI Mode, AI Overviews and Gemini, not Discover and News.", "content_text": "Google has redesigned the Reader Revenue Manager home page. Its new text says content from subscribers and contributors can appear in a \"From your subscriptions\" distinction on Google Search, AI Mode, AI Overviews and Gemini, where the old text named a panel on Search, Discover and News products. Google announced no change separately.\n\n- Google's Reader Revenue Manager home page now says subscription content can appear in a \"From your subscriptions\" distinction on Google Search, AI Mode, AI Overviews and Gemini.\n- An Internet Archive snapshot from 6 October 2026 shows the old wording, a \"From Your Subscriptions\" panel on Google Search, Discover and News products.\n- The new Advanced features section says readers link their publisher subscription account to their Google account to see your content across Google.\n- The page gives no eligibility rules, no traffic data and no launch dates for these surfaces, and it says content \"can\" appear, not that it will.\n- Google's Search Central documentation updates page has no entry about Reader Revenue Manager, and Search Engine Roundtable's report is how the change surfaced.", "date_published": "2026-10-09T00:00:00.000Z", "date_modified": "2026-10-09T00:00:00.000Z", "tags": [ "news", "ai-search", "content-strategy" ] }, { "id": "https://surfil.com/news/google-ugc-fresh-data-program", "url": "https://surfil.com/news/google-ugc-fresh-data-program", "title": "Google adds UGC Fresh Data Program for large platforms", "summary": "Google's UGC Fresh Data Program lets approved platforms push user-generated content to Search outside crawling and the Indexing API.", "content_text": "Google has added a Search Central page for the UGC Fresh Data Program, a dedicated pipeline where approved platforms send user-generated content and interaction signals to Google, separate from the Indexing API and normal crawling. Participation is limited to large, public, moderated platforms, and Google does not guarantee that content will appear in Search.\n\n- Google's new documentation describes a dedicated data ingestion pipeline through which approved platforms send user-generated content and interaction signals to Google.\n- Google says the program is independent of the Google Indexing API and of traditional organic crawling for general web content.\n- Eligibility covers content focus, volume and audience, public accessibility, technical readiness, moderation and freshness, so most sites do not qualify.\n- Applying means submitting a form, Google promises a response within 6 to 8 weeks, and acceptance is not guaranteed.\n- Google states that the program does not guarantee that content will appear in Search, and it does not change how ordinary sites are crawled.", "date_published": "2026-10-08T00:00:00.000Z", "date_modified": "2026-10-08T00:00:00.000Z", "tags": [ "news", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/news/manual-action-requests-rise-reports", "url": "https://surfil.com/news/manual-action-requests-rise-reports", "title": "SEOs report more manual actions as Google stays silent", "summary": "Marie Haynes and Daniel Foley Carter say manual action cases are rising. It is anecdotal, with no Google data, and the spam update finished the same day.", "content_text": "Two SEO consultants, Marie Haynes and Daniel Foley Carter, say they are seeing more Google manual actions, mainly for thin content, pure spam and scaled content. Search Engine Roundtable reported it on 8 October 2026. These are anecdotes from consultants, Google has published no data or statement, and the reports do not show why.\n\n- Marie Haynes said on X on 7 October 2026 that she has had a big increase in requests for help with manual actions, mostly thin content, pure spam and scaled content.\n- She told Search Engine Roundtable the volume went from essentially no requests to several per week, which is not nearly the pre-Penguin level.\n- Daniel Foley Carter posted that he is also seeing more manual actions, and Barry Schwartz found more forum complaints but not a big spike.\n- Google has published no count of manual actions and has made no statement, so the reports are anecdotes and say nothing about cause.\n- A manual action is issued by a human reviewer, shows in the Search Console Manual actions report and message center, and is lifted only after a reconsideration request.", "date_published": "2026-10-08T00:00:00.000Z", "date_modified": "2026-10-08T00:00:00.000Z", "tags": [ "news", "core-updates", "search-console" ] }, { "id": "https://surfil.com/news/merchant-center-frequent-changes-disapproval", "url": "https://surfil.com/news/merchant-center-frequent-changes-disapproval", "title": "Merchant Center may disapprove products that change often", "summary": "Google's Merchant Center help page now warns that products whose price or stock changes more than daily may be disapproved instead of auto-updated.", "content_text": "Google's Merchant Center Help page on automatic updates now warns that automations may not work for products whose price or availability changes often, for example more than once per day. A product with a detected mismatch may be disapproved instead of updated, and Google may stop updates when total mismatches grow too large.\n\n- Google's help page says automations may not work for products with frequent price or availability changes, giving more than once per day as an example.\n- The page adds that a product may be disapproved instead of updated on a detected mismatch, and that updates may stop if total mismatches become too large.\n- The 17 March 2026 archived copy of the page does not contain these sentences, so they are new since then.\n- Hana Kobzova of PPC News Feed dates the revision to 27 September 2026, while the help page itself shows no revision date.\n- The page does not give a mismatch threshold or say how a disapproval clears, and it still tells merchants to keep providing accurate product data frequently.", "date_published": "2026-10-08T00:00:00.000Z", "date_modified": "2026-10-08T00:00:00.000Z", "tags": [ "news", "ecommerce-seo" ] }, { "id": "https://surfil.com/news/september-2026-spam-update-completed", "url": "https://surfil.com/news/september-2026-spam-update-completed", "title": "The September 2026 spam update finished after 13 days", "summary": "Google's September 2026 spam update finished on 8 October after 13 days and 16 hours, the longest of its 2026 spam updates. What it means for your data.", "content_text": "Google's September 2026 spam update finished rolling out on 8 October 2026, after 13 days and 16 hours, according to the Search Status Dashboard. That is the longest of the four spam updates Google ran this year, against 19 hours 30 minutes to 2 days 16 hours for the other three. Google has not said what this one targeted.\n\n- The Search Status Dashboard shows the September 2026 spam update beginning on 24 September at 09:15 PDT and ending on 8 October at 01:00 PDT.\n- Its 13 days and 16 hours compare with 19 hours and 30 minutes in March, 2 days and 1 hour in June, and 2 days and 16 hours in August.\n- Google said at launch that the rollout might take up to two weeks, and it used nearly all of that time.\n- Neither the dashboard nor Google's spam updates page says what this update addressed or why it ran longer than the earlier ones.\n- With the rollout complete, 24 September to 8 October is a fixed window for comparing Search Console data, and a dip inside it is not proof the update caused it.", "date_published": "2026-10-08T00:00:00.000Z", "date_modified": "2026-10-08T00:00:00.000Z", "tags": [ "news", "core-updates" ] }, { "id": "https://surfil.com/news/google-local-panel-whatsapp-button", "url": "https://surfil.com/news/google-local-panel-whatsapp-button", "title": "Google's local panel now shows a WhatsApp button", "summary": "Google is showing a WhatsApp button in local panel results, beside or instead of chat. Google has made no statement and no rollout detail exists.", "content_text": "Google is showing a WhatsApp button in local panel results in Search, in place of chat or beside the call, directions, website and share buttons. Sachin Patel spotted it and Barry Schwartz replicated it, so it may be rolling out. Google has made no statement about it, and rollout details are unknown.\n\n- Google is showing a WhatsApp button in local panel results, in place of chat or in addition to the other action buttons.\n- Sachin Patel posted the sighting on X and at Search Engine Watch, and Barry Schwartz replicated it, so it may be rolled out fully.\n- Schwartz recalls Google asking businesses for WhatsApp numbers a couple of years ago, and adding numbers itself for some a few months ago.\n- Google has made no statement about this display, and the sources give no rollout details.\n- Google's help page says chat options are available in select regions only, and that a profile with both WhatsApp and text message shows only the text message option.", "date_published": "2026-10-07T00:00:00.000Z", "date_modified": "2026-10-07T00:00:00.000Z", "tags": [ "news", "local-seo", "serp-features" ] }, { "id": "https://surfil.com/news/google-synthid-detector-opens-to-public", "url": "https://surfil.com/news/google-synthid-detector-opens-to-public", "title": "Google opens its SynthID detector to everyone", "summary": "Google's SynthID Detector is now open to anyone, globally in English. It checks images, video and audio for watermarks, and nothing ties it to rankings.", "content_text": "Google opened its SynthID Detector to everyone on 7 October 2026, available globally in English. It checks image, video and audio files for SynthID watermarks from Google and partners including OpenAI, NVIDIA and Kakao. It is not a general AI detector, and neither Google's post nor Search Engine Journal ties it to rankings.\n\n- Google opened SynthID Detector to anyone on 7 October 2026, globally in English, after an early version last year for media professionals.\n- The tool checks image, video and audio files for SynthID watermarks from Google and partners including OpenAI, NVIDIA and Kakao, with Apple listed as coming soon.\n- Search Engine Journal reports the site says it is not a general AI detector, so a file with no watermark may still have been made with AI.\n- A positive result reads made or edited with Google AI, so a touched-up photo and one created from scratch can get the same label.\n- Neither Google's post nor the report links the detector to Google Search rankings, and the post lists no text files among the supported formats.", "date_published": "2026-10-07T00:00:00.000Z", "date_modified": "2026-10-07T00:00:00.000Z", "tags": [ "news", "ai-search", "content-strategy" ] }, { "id": "https://surfil.com/news/healthcare-review-snippets-drop-schema-app", "url": "https://surfil.com/news/healthcare-review-snippets-drop-schema-app", "title": "Healthcare review snippets have dropped, Schema App reports", "summary": "Schema App says review snippets fell for its healthcare customers since late May. Google has announced no change and its docs do not mention healthcare.", "content_text": "Schema App says its healthcare customers lost Review Snippet visibility in Google Search, with a drop since late May and another in early August for some. Search Engine Roundtable reports Google seems to no longer show review rich results for healthcare pages. Google has announced no change, and its review snippet documentation does not mention healthcare.\n\n- Schema App says healthcare customers saw Review Snippet visibility fall since late May, with another drop in early August for some organizations.\n- Schema App tested aggregateRating connections, alternative property paths and multi-typing healthcare entities with Product, and none restored the rich result.\n- Search Engine Roundtable reports on 7 October 2026 that Google seems to no longer show review rich results for healthcare pages.\n- Google has not announced a change, and neither its review snippet documentation nor its documentation updates log mentions healthcare.\n- Schema App says the pages it analyzed remain eligible and it has not tied the decline to a drop in overall organic traffic.", "date_published": "2026-10-07T00:00:00.000Z", "date_modified": "2026-10-07T00:00:00.000Z", "tags": [ "news", "structured-data", "serp-features" ] }, { "id": "https://surfil.com/news/merchant-center-ucp-integration-hub", "url": "https://surfil.com/news/merchant-center-ucp-integration-hub", "title": "Merchant Center adds a UCP integration hub for US merchants", "summary": "Google's Merchant Center is rolling out a UCP integration hub in the US, with checkout buttons for AI Mode and Gemini behind an interest form.", "content_text": "Google Merchant Center is gradually rolling out a UCP integration hub in the US, where merchants onboard to the Universal Commerce Protocol program. Per Google's help page, UCP lets a store add a checkout button to eligible products in AI Mode and Gemini. Merchants must finish the technical work and submit an interest form first.\n\n- Google's help page says the Merchant Center UCP integration hub is rolling out gradually in the US, with Australia and Canada to follow next year.\n- Per the help page, UCP is an open standard for agentic commerce, and integrating it lets a merchant add a checkout button on eligible product listings in AI Mode and on Gemini.\n- Merchants must complete the technical implementation in the UCP developer guide and fill out an interest form before they can be selected for onboarding.\n- The hub lets merchants pick UCP capabilities, which the help page lists as cart transfer to the merchant site, native checkout on Google and identity linking.\n- Merchants who sell through a third-party platform are told to ask that provider about its UCP plans rather than onboard on their own.", "date_published": "2026-10-07T00:00:00.000Z", "date_modified": "2026-10-07T00:00:00.000Z", "tags": [ "news", "ecommerce-seo", "ai-search" ] }, { "id": "https://surfil.com/news/chatgpt-product-result-labels-test", "url": "https://surfil.com/news/chatgpt-product-result-labels-test", "title": "ChatGPT tests green and gray labels on product results", "summary": "OpenAI is testing labels such as best all-rounder and best for beginners on ChatGPT product results, in green and gray. OpenAI has not commented.", "content_text": "ChatGPT is testing labels on its product results that describe each product, such as best all-rounder, best for beginners and best cushioning. Brodie Clark spotted a green label for the best all-rounder and gray labels for other criteria. OpenAI has made no statement and the sources give no rollout details, so this is an observed test.\n\n- ChatGPT is testing labels appended to product results that add information about the product, such as best all-rounder, best for beginners and best cushioning.\n- The test shows a green label for the best all-rounder product and gray labels for the best product on other criteria.\n- Brodie Clark spotted it and posted about it on X and in his notes, and SERP Alerts posted it too.\n- OpenAI has made no statement and the sources give no rollout information, so this is an observed test and not a launch.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "ai-search", "ecommerce-seo" ] }, { "id": "https://surfil.com/news/google-maps-policy-merchants-listings-posts", "url": "https://surfil.com/news/google-maps-policy-merchants-listings-posts", "title": "Google Maps policy now limits promotions and pricing in listings", "summary": "Google expanded the advertising and solicitation rules for Maps and Business Profiles: no time-bound promotions, pricing or calls-to-action in the listing.", "content_text": "Google expanded the advertising and solicitation section of its Maps and Business Profile content policy. A new rule for merchants bars time-bound promotions, pricing and direct calls-to-action from the listing itself, and bars unrelated emails, phone numbers and links from posts. Offers and updates that may read as promotional are otherwise allowed in posts.\n\n- Google's policy now says contributions to Maps should reflect genuine experiences or define a business's permanent identity, not advertising or solicitation.\n- In the listing, promotional or commercial content with time-bound promotions, pricing or direct calls-to-action is not allowed.\n- In posts, email addresses, phone numbers, social media links and links to other websites unrelated to your business are not allowed.\n- Offers and updates that may be seen as promotional are otherwise permitted in posts, so posts remain the place for them.\n- Search Engine Roundtable credits Hiroko Imai with spotting the change and sharing before and after screenshots on LinkedIn.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "local-seo" ] }, { "id": "https://surfil.com/news/mueller-large-product-aggregator-indexing", "url": "https://surfil.com/news/mueller-large-product-aggregator-indexing", "title": "Mueller: why a large aggregator is crawled but barely indexed", "summary": "John Mueller told a site owner that URLs indexed from a large product aggregator redirect to the homepage, and that such sites are hard to justify.", "content_text": "John Mueller told a site owner that a large product aggregator is hard to make and maintain well, and that many of the site's indexed URLs redirect to the homepage. Search Engine Roundtable reports the asker said Googlebot crawled the site 1.17M times in a few weeks while Search Console showed only 4 indexed URLs.\n\n- Mueller said many of the asker's indexed URLs redirect to the homepage, which he called a big technical issue still to resolve.\n- He described a large product aggregator as not trivial to make and maintain well, because it ends up with a ton of URLs of apparently programmatic content.\n- He said convincing search engines and users that such a site is worthwhile will be hard.\n- Heavy crawling is not evidence of indexing, since the asker saw a spike in crawl activity while the page indexing report listed only four indexed pages.\n- The wording below is as Search Engine Roundtable quotes it, because we could not open Mueller's Bluesky post.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/news/mueller-sitemaps-ai-crawlers-couldnt-fetch", "url": "https://surfil.com/news/mueller-sitemaps-ai-crawlers-couldnt-fetch", "title": "Mueller: AI crawlers read sitemaps, Couldn't fetch explained", "summary": "Google's John Mueller says AI crawlers can't submit sitemaps and that Couldn't fetch on a valid sitemap often comes down to host load or crawl demand.", "content_text": "Google's John Mueller says AI training crawlers usually offer no way to submit a sitemap, so a default name like sitemap.xml or an RSS feed is how they find a site, according to Search Engine Journal. He also said a valid sitemap can show Couldn't fetch in Search Console because of host load or low crawl demand.\n\n- Mueller said AI training crawlers usually have no console to submit a sitemap, so sites that want them to find content should keep sitemap.xml as the name or publish RSS feeds.\n- Mueller said he has seen an AI crawler fetch his sitemap and RSS files in his server logs, but he did not name the crawlers or say what they do with the files.\n- A private sitemap with an unusual name, left out of robots.txt and submitted to Google directly stays hidden from other systems, and Bing would need its own submission.\n- Search Console can report Couldn't fetch for a valid, public sitemap, and Mueller gave two causes, which are Google's host load and crawl demand tied to perceived site quality.\n- Search Engine Journal notes both causes sit outside the sitemap file, so a clean validation does not rule the error out.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/news/openai-eu-chatgpt-text-watermark", "url": "https://surfil.com/news/openai-eu-chatgpt-text-watermark", "title": "OpenAI will watermark ChatGPT text in the EU", "summary": "OpenAI adds an invisible watermark to ChatGPT and Codex text in the EU, with an API opt-in worldwide. Edits weaken it, and Google does not use it.", "content_text": "OpenAI will add an invisible watermark to qualifying ChatGPT and Codex text in the EU over the coming weeks, according to Search Engine Journal. API users worldwide can opt in. Editing weakens the signal, the detector is limited to approved researchers, and nothing in the report links it to Google rankings.\n\n- OpenAI will watermark qualifying ChatGPT and Codex text in the EU over the coming weeks across all plans, while API users worldwide can opt in on select models, off by default.\n- In OpenAI's own tests on 400-token English passages, swapping 25% of the words for synonyms cut detection from about 92% to 17%.\n- At launch the detector is open only to approved researchers and expert organizations, so ordinary publishers cannot scan their own pages.\n- Neither the report nor OpenAI's update says the watermark is a Google signal, and Google's published guidance on AI content is that quality matters.\n- The update does not say which models the API opt-in covers or whether EU ChatGPT users can turn the watermark off.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "ai-search", "content-strategy" ] }, { "id": "https://surfil.com/news/search-console-ai-control-child-property-override", "url": "https://surfil.com/news/search-console-ai-control-child-property-override", "title": "Search Console flags child property AI control overrides", "summary": "Google's help page now says top-level domain owners see a notice on the Search generative AI control page when a child property overrides the parent.", "content_text": "Google added a line to the Search generative AI control help page saying owners of a top-level domain property will see a notification on that page when one of their child properties overrides the parent setting. Search Engine Roundtable spotted the change, and had no screenshot of the notification itself.\n\n- Google's help page now says owners of the top-level domain property see a notification on the Search generative AI control page when a child property overrides the parent setting.\n- A property inherits its control from its closest parent that has stopped inheriting, and falls back to the top-level domain property when none has.\n- Inherit control from parent is the default for any property that has a parent, and any child owner can override it at any time.\n- The control covers AI Overviews, AI Mode and generative AI features in Discover, and Google says it does not affect AI training or ranking in other parts of Search.\n- Google says an exclusion generally takes a few days to apply, and Search Engine Roundtable had no screenshot of what the new notification looks like.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "search-console", "ai-search" ] }, { "id": "https://surfil.com/news/search-console-multi-country-filter", "url": "https://surfil.com/news/search-console-multi-country-filter", "title": "Search Console now groups multiple countries in one filter", "summary": "Search Console's performance report now lets you tick several countries and see their combined Google Search traffic, per two trade outlets.", "content_text": "Google has updated the Search Console performance report so you can select several countries in the country filter and view their combined Google Search traffic. Search Engine Roundtable and Search Engine Journal both report the change, quoting Google's LinkedIn announcement, which we could not open ourselves. Search Engine Journal says device filters changed too.\n\n- The performance report's country filter now accepts several countries at once and shows their combined Google Search traffic.\n- Search Engine Roundtable's example ticks the United States and Canada, then clicks update to group them.\n- Search Engine Journal reports that the update covers device filters as well as countries.\n- Google's LinkedIn post could not be opened by us, so its wording here is as quoted by Search Engine Roundtable.\n- Search Engine Roundtable adds that it would like the same click data in the AI performance reports.", "date_published": "2026-10-06T00:00:00.000Z", "date_modified": "2026-10-06T00:00:00.000Z", "tags": [ "news", "search-console", "analytics" ] }, { "id": "https://surfil.com/news/google-spam-updates-ai-scaled-content-deep-dive", "url": "https://surfil.com/news/google-spam-updates-ai-scaled-content-deep-dive", "title": "Google ties more spam updates to a flood of new content", "summary": "At Search Central Live, Google said spam updates target much new content, AI helps catch it, and scaled content now worries it more than link spam.", "content_text": "Google speakers at Search Central Live Deep Dive in Barcelona said a high percentage of content is spam, that spam updates target it, and that there have been more of them because there is more new content. Per a ROAST recap, Google uses AI to catch more spam, and Gary Illyes called scaled content a bigger problem than link spam.\n\n- Per the ROAST recap, Google said a high percentage of content is spam, that spam updates target it, and that there is more new content, hence more updates.\n- The recap says the spam team now uses AI to help catch more spam, and Google's own spam page describes machine learning and large language models in its spam-fighting systems.\n- Gary Illyes said Google filters out 40 billion spam pages a day, a figure Google words as found, not removed, and Search Engine Roundtable notes is not new.\n- Core updates target content at page level, not domain level, so some pages rise while others fall, and Illyes called scaled content a bigger problem than link spam.\n- Google's quality guidance, as recapped, defines quality through effort, originality, talent or skill and accuracy, and says low quality content can be written by a human or by AI.", "date_published": "2026-10-05T00:00:00.000Z", "date_modified": "2026-10-05T00:00:00.000Z", "tags": [ "news", "core-updates", "content-strategy" ] }, { "id": "https://surfil.com/news/search-console-dark-mode-slide-was-a-joke", "url": "https://surfil.com/news/search-console-dark-mode-slide-was-a-joke", "title": "Search Console dark mode was a joke, per John Mueller", "summary": "A dark mode mock-up shown at Search Central Live in Barcelona was first reported as coming, then described as a joke. Here is what each source says.", "content_text": "Google is not building dark mode for Search Console, according to an update to Search Engine Roundtable's report. The site first said Google would soon support it after a Barcelona conference slide, then added that John Mueller confirmed the slide was a joke. A conference recap calls it a mock-up. No release date was given.\n\n- Search Engine Roundtable first reported that Google will soon support dark mode in Search Console, based on slides from the Search Central Live Deep Dive Europe event in Barcelona.\n- The same article later added an update saying the slide was a joke and that Google is not working on dark mode for Search Console.\n- Search Engine Roundtable says John Mueller confirmed it was a joke, so the report no longer supports a dark mode launch.\n- ROAST's recap of the same session describes a mock-up shown at the end that \"isn't real\", and gives no date.\n- The session also covered Groups, AI-powered configuration and Generative AI reporting, where the team gave no dates for more AI search data.", "date_published": "2026-10-05T00:00:00.000Z", "date_modified": "2026-10-05T00:00:00.000Z", "tags": [ "news", "search-console" ] }, { "id": "https://surfil.com/blog/google-crawling-indexing-serving-timelines-2026", "url": "https://surfil.com/blog/google-crawling-indexing-serving-timelines-2026", "title": "How long Google takes to crawl, index and update results", "summary": "All 23 typical and slowest times Gary Illyes showed at Search Central Live Barcelona, as recapped by ROAST, plus caveats and Google's own docs.", "content_text": "Google's Gary Illyes showed typical and slowest times for 23 crawling, indexing and serving processes at Search Central Live Deep Dive in Barcelona on 2 October 2026. Per the ROAST recap, a typical site move takes 1-3 months, a title update 1-2 days and core update recovery 3-6 months. The figures are Google's internal analysis, not published documentation.\n\n- Per the ROAST recap, Gary Illyes showed a typical and a slowest time for 23 processes across crawling, indexing and serving, based on Google's internal analysis.\n- The recap lists a typical site move at 1-3 months and core update recovery at 3-6 months, with slowest cases of 6 months to 1 year or more.\n- Five of the 23 slowest times include the word never, and three of those add quality in brackets, so a fast technical fix cannot rescue a page Google does not value.\n- Search Engine Journal notes that typical is undefined and that sample size and measurement period are not reported, so the figures are ranges, not deadlines.\n- Illyes cautioned that the processes are linked, so a delay in crawling carries into indexing and serving.", "date_published": "2026-10-05T00:00:00.000Z", "date_modified": "2026-10-05T00:00:00.000Z", "tags": [ "reports", "indexing", "technical-seo" ] }, { "id": "https://surfil.com/tools/rss-feed-validator", "url": "https://surfil.com/tools/rss-feed-validator", "title": "RSS feed validator: check RSS 2.0 and Atom feeds", "summary": "RSS feed validator: paste feed XML to check RSS 2.0 and Atom 1.0 for parse errors, required elements, dates, guids, enclosures and URLs. Runs in-browser.", "content_text": "This RSS feed validator parses pasted feed XML in your browser, detects RSS 2.0, Atom 1.0 or neither, and reports parse errors first. It then checks the required channel and feed elements, item and entry rules, RFC 822 and RFC 3339 dates, duplicate guids and ids, enclosures, and relative URLs.\n\n- This tool parses XML in the browser with DOMParser, detects RSS 2.0 or Atom 1.0 from the root element, and reports real parse errors first.\n- An RSS 2.0 channel needs a title, a link and a description, and every item needs a title or a description, per the RSS 2.0 specification.\n- An Atom feed and each of its entries need exactly one id, title and updated, and an entry without content needs an alternate link, per RFC 4287.\n- Google accepts RSS 2.0 and Atom 1.0 feeds as sitemaps, and says such a feed only provides information on recent URLs.\n- This validator checks structure only, so a clean pass does not confirm the listed pages exist, load or can be indexed.", "date_published": "2026-10-04T00:00:00.000Z", "date_modified": "2026-10-04T00:00:00.000Z", "tags": [ "tools", "technical-seo", "indexing" ] }, { "id": "https://surfil.com/tools/utm-link-builder", "url": "https://surfil.com/tools/utm-link-builder", "title": "UTM link builder", "summary": "UTM link builder: add utm_source, utm_medium and utm_campaign to any URL, keep its query string and fragment, and copy the link. Free, in your browser.", "content_text": "This UTM link builder adds campaign parameters to a URL so Google Analytics can tell where visits come from. Source, medium and campaign name are required, term and content are optional. It keeps any existing query string, replaces a utm_ parameter of the same name, leaves the fragment last, and lowercases values by default because Analytics is case sensitive.\n\n- Google Analytics Help says to always use utm_source, utm_medium and utm_campaign.\n- utm_term and utm_content are optional, and utm_content is for telling similar links in the same ad apart.\n- Analytics is case sensitive, so utm_source=google and utm_source=Google are different values.\n- Google's own best practice is to use lowercase as a standard practice.\n- Parameters follow the base URL after a question mark and are separated from each other with an ampersand.", "date_published": "2026-10-03T00:00:00.000Z", "date_modified": "2026-10-03T00:00:00.000Z", "tags": [ "tools", "analytics" ] }, { "id": "https://surfil.com/news/bing-recommended-by-products-test", "url": "https://surfil.com/news/bing-recommended-by-products-test", "title": "Bing tests a Recommended by label on product results", "summary": "Bing is testing a Recommended by section that names a company beside product thumbnails, with no link to it. Microsoft has not commented.", "content_text": "Bing is testing a section labeled Recommended by, followed by a company name, above product thumbnails that link to retailers. Sachin Patel spotted it, and Barry Schwartz replicated it in some browsers. There is no way to verify the recommendation, and Microsoft has not commented or shared any rollout plans.\n\n- Bing is testing a Recommended by section that names the company making the recommendations.\n- The product thumbnails link to various retailers, not to the recommender, and no source of the recommendation was seen.\n- Sachin Patel posted the sighting on X and at Search Engine Watch, and Barry Schwartz replicated it in some browsers.\n- Microsoft has made no statement and the source gives no rollout information, so this is an observed test.", "date_published": "2026-10-02T00:00:00.000Z", "date_modified": "2026-10-02T00:00:00.000Z", "tags": [ "news", "serp-features", "ecommerce-seo" ] }, { "id": "https://surfil.com/news/google-discover-eligibility-gate-deep-dive", "url": "https://surfil.com/news/google-discover-eligibility-gate-deep-dive", "title": "Google Discover eligibility gate: what Google's analyst said", "summary": "Google's Discover trust and safety analyst described a three-way eligibility gate and image rules at Search Central Live Barcelona, per a ROAST recap.", "content_text": "Google's Discover trust and safety analyst, Eric Murillo, said Discover runs on the normal Google index stack with no special bot, according to a ROAST recap. Content is checked for violations, then judged on E-E-A-T, and only safe, trusted content becomes eligible. Large images then drive performance. These are the recap author's notes, not Google's slides.\n\n- Per the recap, Eric Murillo said there is no special Discover bot and everything runs on the normal Google index stack.\n- Every piece of content goes through an eligibility gate that filters violations, filters low quality judged on E-E-A-T, and passes safe, trusted content as Discover eligible.\n- The recap's image advice is 1,200px wide or more, 300,000 or more total pixels, 16:9 and max-image-preview:large, which matches Google's Discover help page.\n- Murillo reportedly said AI slop is increasing and that user feedback is what new spam filters are built from.\n- A manual action against your site shows in Search Console, as both the recap and Google's Discover help page say.", "date_published": "2026-10-02T00:00:00.000Z", "date_modified": "2026-10-02T00:00:00.000Z", "tags": [ "news", "serp-features", "content-strategy" ] }, { "id": "https://surfil.com/news/google-fake-author-profiles-warning", "url": "https://surfil.com/news/google-fake-author-profiles-warning", "title": "Google warns against fake author profiles in its guidance", "summary": "Google's helpful content page now calls fabricated creator profiles deception: AI headshots, made-up names and false credentials.", "content_text": "Google's helpful content page now warns against deceptive authorship information. In the version dated 1 October 2026, it calls fabricated creator profiles, such as AI-generated headshots, made-up names or false credentials, a form of deception. It says deception makes a page untrustworthy to users and to Google's automated quality systems.\n\n- The new paragraph sits in the \"Who (created the content)\" section, right after the advice to add accurate authorship information such as bylines.\n- Google names AI-generated headshots, made-up names and false credentials as examples of fabricated creator profiles.\n- Google says any form of deception makes a page untrustworthy to both users and its automated quality systems, and is a signal of a low-quality page.\n- Search Engine Journal reports that the rater guidelines already treat fake profiles as deception, and that the paragraph does not say how automated systems detect them.", "date_published": "2026-10-02T00:00:00.000Z", "date_modified": "2026-10-02T00:00:00.000Z", "tags": [ "news", "content-strategy", "core-updates" ] }, { "id": "https://surfil.com/news/google-helpful-content-main-content-section", "url": "https://surfil.com/news/google-helpful-content-main-content-section", "title": "Google adds a main content section to its helpful content page", "summary": "Google's people-first content page now defines main content and lists four attributes raters assess: effort, originality, talent or skill, and accuracy.", "content_text": "Google's helpful content page now has a Main content section, reported on 2 October 2026. It defines main content as any part of a page that directly helps it achieve its purpose, and lists four attributes raters assess: effort, originality, talent or skill, and accuracy. Google repeats that rater data is not used directly in ranking.\n\n- The page now says one of the most critical factors in assessing page quality is the quality of the main content, beyond E-E-A-T.\n- Google defines main content as any part of the page that directly helps it achieve its purpose, from text and media to calculators, reviews and headings.\n- Raters are trained to assess four attributes of main content: effort, originality, talent or skill, and accuracy.\n- Google says crediting other sources does not replace original effort, and names generating pages from feeds or bulk unsupervised generative AI text as little to no effort.\n- The page still says raters do not control rankings, rater data is not used directly in ranking algorithms, and E-E-A-T is not a specific ranking factor.", "date_published": "2026-10-02T00:00:00.000Z", "date_modified": "2026-10-02T00:00:00.000Z", "tags": [ "news", "content-strategy", "core-updates" ] }, { "id": "https://surfil.com/news/google-query-understanding-synonyms-fan-out-deep-dive", "url": "https://surfil.com/news/google-query-understanding-synonyms-fan-out-deep-dive", "title": "How Google turns a query into results, per a Barcelona recap", "summary": "A ROAST recap of Mueller and Illyes in Barcelona: synonym expansion, AI fan-out queries missing from Search Console, and retrieval before ranking.", "content_text": "At Google's Search Central Live Deep Dive in Barcelona, John Mueller and Gary Illyes described how a query becomes results, according to a ROAST attendee recap. Synonym expansion works like an OR search, AI search adds LLM fan-out queries that return to the search engine, and retrieval comes before ranking.\n\n- Per the recap, Mueller called synonym expansion a very important ranking system that works much like searching with the OR operator.\n- In AI search, an LLM creates fan-out queries that go back to the search engine, and the results ground the answer.\n- The recap says Google's fan-out queries do not appear in Search Console because they do not come from users.\n- Illyes described retrieval as happening before ranking, with language and country as important signals and quality deciding the final result.\n- The recap says rich results are generally driven by structured data, while AI results do not need structured data.", "date_published": "2026-10-02T00:00:00.000Z", "date_modified": "2026-10-02T00:00:00.000Z", "tags": [ "news", "ai-search", "content-strategy" ] } ] }