INDEPENDENT MINDS. SHARED AMBITION.

Newsfeed

Find the original sources, practical reading and people behind the collective. These links open the full public pages.

1,808 pages · Page 4 of 19

  • Bing tests a new way to show prices in ads

    Sachin Patel has spotted Bing testing a new way to show prices in search ads. The test surfaces pricing as a more prominent visual element within the standard text ad format: an evolution beyond the existing Price extension surface, which has been live for years.

  • Bing tests double-row product ad carousel and near-transparent sponsored labels

    Sachin Patel has spotted two simultaneous Microsoft Advertising display tests on Bing. The first shows a significantly enlarged sponsored product section in a double-row carousel alongside organic results: a substantial increase in ad inventory density on the Shopping SERP.

  • Bing tests large image assets in Search ads

    Microsoft Bing is testing large images in Search ads that would substantially increase the visual footprint of sponsored results, with early coverage suggesting a material upside in CTR. For you, this means Bing SERPs may get more visual, so prepare image assets even if Microsoft is a secondary channel for you.

  • Bing updates Shopping ads design

    Microsoft has updated the visual design of its Shopping ads with a refreshed layout. The update was spotted in the wild, though Microsoft has not officially announced the specific changes.

  • Bing Webmaster Tools adds page-level citation data for AI grounding queries

    Microsoft has updated the AI Performance Dashboard in Bing Webmaster Tools to surface bidirectional citation insights: advertisers and SEOs can now see which specific pages are being cited for a given grounding query, and which grounding queries are driving citations to a specific page. Aleyda Solís flagged the update as a direct response to community feedback, tagging the Bing team by name.

  • Book Online buttons extend into Search and Performance Max

    Reserve with Google appointment links can now appear as Book Online actions in eligible Search and Performance Max ads, using the same partner connection already published for local surfaces. The advertiser pays for the ad click, while an encrypted token links the visit to a booking and the “Appointments booked” conversion must be included in campaign goals for bidding; healthcare is excluded from the current setup.

  • Bram Van der Hallen reports four new Meta Ads updates

    Bram Van der Hallen identified four updates rolling out across Meta Ads this week. Instagram has been added to the Manus AI connector, allowing advertisers to manage and analyse Instagram content: including posts, stories, and Reels: through the Manus interface, and to configure automations and schedules for publications.

  • Brand Lift Studies adds a new Association metric

    Thomas Eccel has flagged a new metric appearing in Google Ads Brand Lift Studies, called Association. The metric lets advertisers select a concept, category, or attribute and measure how strongly users associate the brand with that concept after exposure to the campaign: a measurement type previously available only through bespoke brand-tracking research with separate vendors.

  • Brand Lift studies expand with an Enhanced option.

    Google Ads has expanded the availability of Brand Lift studies and added an Enhanced option, as Thomas Eccel flagged. The Enhanced study carries a lower reported lift threshold but requires a larger budget than the Standard version.

  • Brand logos are now visible in Google Shopping results

    Google Shopping has begun displaying brand logos next to merchant names: a small visual change with potentially significant implications for brand advertisers. The logo serves as an immediate trust signal, distinguishing established brands from generic resellers in a results page where products often appear nearly identical aside from price.

  • Brand names appear in YouTube product Ads + Shorts Shopping ads

    Arpan Banerjee discovered that brand names are now appearing in YouTube product Ads: including the YouTube Shorts Shopping ad surface: a finding flagged by Adriaan Dekker on LinkedIn. Where YouTube product Ads previously displayed product images, price, and store name, the brand name now appears as an additional surface element.

  • Branded queries filter now available to all eligible sites in Search Console

    Google Search Console has rolled out the branded queries filter to all eligible sites. The feature automatically differentiates between branded and non-branded queries in the Performance Report, helping analyse traffic from users familiar with your brand versus those discovering you organically.

  • Budgets go weekly, UTMs reach all levels, and vertical policies clarify

    Three ChatGPT Ads changes landed together. For you, this means the platform is maturing fast while shifting spend control, so recheck budget pacing under the weekly model.

  • Build-to-Order availability attribute added for vehicle listings

    Google Merchant Centre has introduced a new "build to order" value for the availability attribute, specifically for vehicle listings. This applies to vehicles that aren't in dealership inventory but can be customised and ordered directly by customers.

  • Bulk ad review now supports specific campaign selection

    Google has added campaign-level filtering to its bulk ad review workflow. Advertisers requesting bulk revision of policy-flagged ads can now select specific campaigns rather than submitting all eligible ads simultaneously: reducing the risk of triggering broad re-review of campaigns that are already serving correctly.

  • Bulk editing arrives for disapproved assets

    Microsoft Advertising now supports bulk editing and appealing of disapproved text and image assets, live in all accounts. For you, this means disapproval cleanup stops being one-by-one, so clear backlogs in a single pass.

  • Business Agent for Leads brings conversational lead qualification to Search ads

    Business Agent for Leads is a new Search ad format that replaces the static lead form with a conversational AI agent built on Gemini. Instead of filling out a form, a prospect researching, for example, an MBA programme can click "Chat" directly within the ad to ask questions, with answers grounded in the advertiser's website content.

  • Business Agent goes live for eligible e-commerce merchants

    The Gemini-powered Business Agent previewed at GML is now live for eligible merchants, flagged by Khushboo Aulakh. The agent sits on a brand's profile in search results and answers product questions in the brand's voice before a customer clicks through: handling queries such as sizing or quantity calculations that would otherwise reach the customer service team.

  • Business data feeds reach Demand Gen campaigns.

    Google is bringing business data feeds to Demand Gen, enabling dynamic ads without a Merchant Centrer account. The capability targets non-retail verticals such as travel, real estate, and automotive, and currently applies to Google Display Network inventory within Demand Gen.

  • Business logo and unit pricing now surfacing in Google Shopping

    Thomas Eccel has spotted two Merchant Centre feed attributes gaining visible placement in Google Shopping listings. Business logos now appear next to the merchant name in sponsored product listings: a trust signal before the click.

  • Business logos can be added directly in Google Ads

    Google Ads now lets advertisers add a business logo during setup for Performance Max and Shopping campaigns, using the Merchant Centre logo as a fallback when no separate asset is supplied. Arpan Banerjee was credited with first bringing the change to light.

  • Campaign Manager 360 API v4 sunsets on February 26, 2026

    Google will shut down Campaign Manager 360 API v4 on February 26, 2026. All v4 requests will fail after that date.

  • Campaign tree gains status dots with inline pause controls

    Meta's campaign, ad set, and ad tree now shows a status dot at each level, with the ability to activate or pause directly from the structure. For you, this means faster housekeeping without opening each level, so use the tree for quick status sweeps.

  • Campaign Type Attribution gives Demand Gen its own full-funnel attribution model

    Campaign Type Attribution is a structural shift for Demand Gen reporting: moving beyond Platform Comparable Conversions (a reporting-only view) into an active attribution model where Demand Gen campaigns can be optimised against the full conversion value they drive, independent of other Google campaign types. Georgi Zayakov flagged it as the most interesting Demand Gen update from GML, noting it makes Demand Gen results "more directly comparable to social platforms": which is clearly where Google wants Demand Gen to sit in the marketing mix.

  • Carousel ads gain a Collage display option.

    Meta's latest ad creation flow lets you show carousel media as a single media ad, with Collage, single image or video, and video slideshow options. The controls sit under Format display options at the ad level and reuse the media from your original carousel.

  • Carrefour becomes the first major European retailer to enable ChatGPT checkout

    Carrefour has become the first major European retailer to make products directly purchasable through ChatGPT, targeting France's 26 million ChatGPT users. The integration enables AI-mediated basket building: ChatGPT handles meal planning, dietary filtering, and product selection on behalf of the shopper without a traditional search results page.

  • Certified US pharmacies can now promote prescription drug subscriptions

    Google updated its billing policy to let certified US Online Pharmacies promote prescription subscriptions, drug-plus-service bundles, and recurring consultation models. Requirements include maintaining certification, using subscription_cost in Merchant Centre, and disclosing all fees.

  • ChatGPT abandons agentic commerce checkout

    OpenAI is abandoning its Instant Checkout feature for ChatGPT shopping, reverting to redirecting users to retailer websites rather than in-app purchases. Staff realised that while users researched products in ChatGPT, they were not using it to actually buy.

  • ChatGPT ad formats evolve toward shopping placements

    Juozas Kaziukėnas has flagged a third iteration of ads inside ChatGPT, with the format evolving to carry a larger image while taking up less horizontal space: a layout that would allow multiple ads side by side, though only one ad shows at a time so far. He notes that more ad formats are coming, specifically for shopping.

  • ChatGPT adds a library for centralised file management across conversations

    OpenAI has launched Library in ChatGPT: a new sidebar tab providing a central location for all files uploaded across conversations, with browsing, sorting, search, and direct attachment from the composer. OpenAI reports that more than 1 billion files are uploaded to ChatGPT each week.

  • ChatGPT Ads adds audiences and launches in Japan and South Korea

    ChatGPT Ads has added audience targeting with advanced bid adjustments, plus custom audiences that let you upload lists of 25,000 or more users to target or exclude, with ad-group-level bid multipliers. OpenAI also added an Overview dashboard, suggested ad drafts built from your site's metadata, a smaller ad-card format, and availability in Japan and South Korea.

  • OpenAI expands the set of customer identifiers for conversion matching

    OpenAI has expanded the set of matching inputs for the ChatGPT Ads Measurement Pixel and Conversions API, including additional customer and geographic fields.

  • ChatGPT Ads expands custom-audience management

    OpenAI’s custom-audience update allows advertisers to add, remove or replace members without rebuilding an audience, mix supported identifiers and build lists beyond five million members.

  • ChatGPT Ads expand to the UK in the coming weeks

    Tariq Mohammed has flagged that ChatGPT Ads are coming to the UK in the next few weeks. OpenAI confirmed on 7 May that UK users on the Free and Go tiers will start seeing ads in the coming weeks: in the shopping, retail, and travel categories first.

  • ChatGPT Ads expands to 31 European markets

    ChatGPT Ads expanded to 31 European markets in August. OpenAI subsequently confirmed that self-service Ads Manager access opened across those markets on 31 August 2026.

  • ChatGPT Ads expands into Brazil and Mexico

    OpenAI’s August update confirms that ChatGPT Ads has launched in Brazil and Mexico. Carlos Romero’s accompanying account of early testing offers practical observations on budget control and campaign focus.

  • OpenAI introduces a ChatGPT Ads Manager plugin

    OpenAI’s latest advertising update includes a plugin for creating, managing and analysing campaigns through conversations in ChatGPT or Codex.

  • OpenAI Ads adds measurement integrations and experiments

    How to assess OpenAI Ads measurement integrations while keeping attribution, experiments and causal lift distinct.

  • ChatGPT Ads pilot expanding to the UK, Japan, South Korea, Brazil, and Mexico

    OpenAI has confirmed that the ChatGPT Ads pilot will expand to the United Kingdom, Japan, South Korea, Brazil, and Mexico in the coming weeks. Benji Shomair (OpenAI) has flagged the geographic expansion alongside a hiring announcement for Client Partner and Customer Success Manager roles in each new market; David Dugan (OpenAI) has confirmed the same expansion with additional team-building context.

  • ChatGPT Ads pre-generates ad variations for review

    ChatGPT Ads now pre-generates ad variations and adds them to existing campaigns after the advertiser reviews and approves them. Approval is currently required before anything goes live.

  • ChatGPT Ads prepare for the EU with an updated privacy policy

    OpenAI's Europe privacy policy, updated in June, sets out how personal data will be handled for advertising to users in the EEA, UK, and Switzerland: the clearest sign yet that ChatGPT Ads' EU launch is close. For you, this means the largest untapped ChatGPT market is being prepared for ads, so plan European tests before launch pricing settles.

  • ChatGPT Ads adds product carousels

    The newsletter reports product carousel ads in ChatGPT, adding a product-led format for eligible advertisers.

  • ChatGPT Ads roll out conversion-optimised CPA campaigns from early June

    OpenAI is rolling out conversion-optimised ads where advertisers only pay when a user buys after clicking, flagged by Juozas Kaziukėnas, with Anthony Higman and Glenn Gabe adding context. ChatGPT Ads launched less than four months ago with CPM ads, added CPC bidding a month ago, introduced a conversion-tracking pixel, and have now started rolling out CPA ads that bid for successful events.

  • ChatGPT self-service advertising expands across more markets

    OpenAI expanded self-service access to ChatGPT Ads Manager across India, Europe, the Middle East and North Africa on 31 August.

  • ChatGPT ads spotted live in the United States

    The Adthena team has identified the first confirmed ads appearing in ChatGPT for signed-in desktop users in the US. The sighting, which Glenn Gabe compared to "seeing Bigfoot in the wild," confirms that OpenAI's advertising pilot has moved from announcement to execution.

  • ChatGPT Ads test now live in the US

    ChatGPT's advertising test is now live in the United States, with users spotting ads appearing in responses. The beta pilot is reportedly closing applications this week with a $200K minimum spend and CPM-based pricing.

  • Launch check: review ChatGPT’s text customisation default

    As self-service ChatGPT advertising expands, practitioners are highlighting a text customisation setting that can generate personalised and translated headlines and descriptions without individual approval.

  • OpenAI adds pacing for campaign total budgets

    OpenAI says ChatGPT Ads campaigns using total budgets now include pacing intended to spread expenditure more evenly across their running period.

  • ChatGPT Ads trial shows low CTR and slow spend pacing

    Ashley Fletcher of Adthena reported further observations from the ongoing ChatGPT Ads trial. One advertiser has spent only 3% of their allocated $250,000 budget.

  • OpenAI permits eligible legal-services ads in the US

    OpenAI’s updated advertising policy permits legal-services advertising in the US when advertisers are licensed in the jurisdiction where the ad is shown.

  • OpenAI previews conversion optimisation, including ad views

    OpenAI plans to introduce a ChatGPT Ads optimisation model that considers conversions following both clicks and ad views, with impression-based billing for that campaign type.

  • ChatGPT appears in Google Ads auction insights

    An advertiser in a niche financial software area found ChatGPT in their Google Ads auction insights, present since May on exact match keywords. It is the clearest sign yet that ChatGPT ads compete in overlapping auctions.

  • ChatGPT builds an e-commerce product feed infrastructure

    ChatGPT is now ingesting product feeds directly from merchants rather than scraping Google Shopping. Shopping results now display source feed names, including Walmart, Target, Home Depot, Chewy, Best Buy, Etsy, and Shopify merchants.

  • ChatGPT carousel reporting adds product-card visibility

    OpenAI’s advertiser briefing adds card impressions and clicks for product-feed carousel ads, as well as product-level reporting via its Insights API.

  • ChatGPT citations shift away from listicles and Reddit

    Tomek Rudzki's comparison of one million prompts before and after the GPT-5.6 rollout found listicles losing half their citation share and comparison pages falling 32.1%, alongside more fan-out searches and more direct-source retrieval. Aleyda Solís separately observed declining Reddit citations across several markets and verticals, with competitors, specialist publications, platforms, and government sites gaining visibility; neither dataset proves the same effect for every brand or model.

  • ChatGPT Deep Research is now powered by GPT-5.2

    Deep research in ChatGPT is now powered by GPT-5.2. You can now connect to apps in ChatGPT, search specific sites, and track real-time progress with follow-ups.

  • ChatGPT desktop tests WebMCP-powered shopping actions

    Juozas Kaziukėnas demonstrated an experimental ChatGPT desktop flow that opened a Shopify store, found a red lipstick, added it to the basket, and navigated to the billing form via structured WebMCP calls. The system used functions exposed by the website rather than scraping pages and simulating every click, while the customer still completed payment details.

  • ChatGPT Ads reporting exposes goal and non-goal events

    How to interpret ChatGPT Ads goal-only conversions and attributed non-goal event reporting.

  • ChatGPT Images 2.5 adds more control over creative edits

    OpenAI has announced ChatGPT Images 2.5, with targeted image comments, sketch references and templates for formats including product photos. The company says it improves reference preservation and consistency across edits, with rollout across ChatGPT, ChatGPT Work and Codex.

  • ChatGPT is buying traffic from the search platform it competes with

    Garrett Sussman shared iPullRank panel data showing ChatGPT as the sixth-largest destination for Google Search clicks in the sample, with about 1.7 million clicks in one week and 4.75% of its Google clicks classified as paid. The analysis cannot reveal OpenAI's total spend, bids, or unclicked auctions, so it measures observed traffic mix rather than budget.

  • ChatGPT keeps gaining US share as AI chat interfaces reach 2% of retail traffic

    Two data points this week underline AI's steady encroachment on discovery. Vladimir Hanzlik noted that ChatGPT continues to add roughly 3 million new US site visitors per month, with its share of US internet users still climbing and no sign of a slowdown despite pressure from Gemini in the consumer market and Claude in the business market.

  • ChatGPT kills dedicated Shopping Research feature after speed complaints

    Juozas Kaziukėnas has reported that OpenAI has quietly removed ChatGPT Shopping Research from the main menu: effectively ending a feature launched only four months ago, which used a version of GPT-5 mini post-trained specifically for shopping tasks. The feature was designed to ask clarifying questions, scrape 10–20× more sources than standard ChatGPT, and produce personalised buyer guides: but typically took two minutes to complete, which OpenAI's new commerce product lead cited as incompatible with user behaviour.

  • ChatGPT launches a native restaurant experience

    ChatGPT now offers a native restaurant experience, returning bookable recommendations in conversation: an early glimpse of super-app behaviour. For you, this means local and hospitality visibility inside ChatGPT is becoming a real surface, so watch how listings and reviews feed its answers.

  • ChatGPT merchant dashboard spotted for product feed management

    Juozas Kaziukėnas shared images of the ChatGPT ecommerce merchant dashboard, currently live only for invited retailers. The interface allows merchants to set up product feeds via SFTP push or URL pull, similar to Google Merchant Centre.

  • ChatGPT policy opens medical, legal and financial advice contexts to ads

    OpenAI has updated the ChatGPT ad policy to remove the categorical block on advertising inside medical, legal, and financial advice contexts. Anthony Higman has flagged the change with its practical implication: previously regulated verticals that were structurally excluded from ChatGPT Ads can now bid, provided their creative and landing pages comply with the platform's broader advertising rules.

  • ChatGPT's browsable product catalog now permanent at the bottom of every shopping conversation

    Juozas Kaziukėnas has reported that ChatGPT's browsable product catalogue: a block of four products with a "Browse All" expansion button: is now appearing at the bottom of all shopping conversations. Kaziukėnas draws a direct parallel to Amazon's "Frequently bought together" section, noting that those slots are now entirely sponsored products, and raises the question of how long ChatGPT will hold off from monetising this placement as it scales ad revenue.

  • ChatGPT source code reveals ad references ahead of launch

    References to "ads shown to the user" have been spotted in the source code of ChatGPT responses, signalling that OpenAI's advertising rollout is imminent. The beta pilot program is reportedly closing applications this week with a $200K minimum spend and CPM-based pricing.

  • ChatGPT tests a separated ad label and site name

    SERP Alert observed a test in which the ad label and advertiser site name appear above standard single-text ads rather than inside the ad unit. The sighting does not establish a full rollout or a performance effect.

  • OpenAI outlines US visual-ad test during image generation

    OpenAI’s planned US visual-ad test: what is known, what remains unspecified and how advertisers can evaluate it.

  • Checkout Links expand to 9 new markets and Shorts in-feed placement

    Checkout Links: the format that routes Demand Gen viewers directly to a merchant's checkout page: is generally available in the US, CA, GB, DE, JP, IN, with rollout coming to AU, KR, ID, MX, AR, FR, IL, CH, PL. Checkout Links have also expanded to YouTube Shorts and in-feed ads in Demand Gen.

  • Chrome announces WebMCP early preview for agentic web interactions

    Google Chrome has announced WebMCP in early preview, a proposed standard that would allow websites to expose structured tools to AI agents. The protocol includes a Declarative API for HTML form actions and an Imperative API for JavaScript-based interactions.

  • Chrome DevTools now explains Google Ads disapprovals and console errors with Gemini AI

    Thomas Eccel has highlighted that Google Chrome DevTools now includes built-in Gemini AI features accessible via Settings → AI Innovations. The tooling includes an "Understand this error" button that breaks down console errors in plain language, AI-generated explanations for failed requests and warnings, code suggestions inside the Sources panel, and AI assistance for CSS styles, network requests, and performance issues.

  • Chrome documents LLMs.txt as part of agentic audits

    Chris Long has flagged a significant Chrome documentation update: Chrome released documentation on its new Agentic Browsing audits, and buried in the spec is a reference to how the audit will check for the LLMs.txt file at the domain root. Crystal Carter was credited as the original finder.

  • Chrome Lighthouse adds an Agentic Browsing audit category

    Google has added an experimental Agentic Browsing category to Chrome's Lighthouse audit tool. Instead of auditing only for human users, the new category tests whether AI agents can read a site, navigate key pages, understand actions and forms, and reliably interact with elements.

  • Circle to Search now supports multi-object search

    Google has updated Circle to Search to explore multiple objects in an image simultaneously. Users can circle an entire outfit to identify all its components, or circle multiple items in any image to get information about each.

  • Clarity publishes five advanced MCP Server queries

    Microsoft Clarity published five advanced query patterns for its MCP Server, including identifying pages with high scroll depth but low conversion rates and diagnosing mobile page friction. The examples are workflow guidance rather than a new advertising feature.

  • Clarity scores whether AI crawlers send visitors back

    Microsoft Clarity's Bot Analytics dashboard adds a Scrape-to-Referral Ratio, showing which AI crawlers extract your content and actually return traffic, and which only take. For you, this means the AI-crawler trade finally has a scoreboard, so use it before deciding which bots to block.

  • Clarity splits AI citations into branded and non-branded

    Microsoft Clarity's Citations dashboard adds segmentation and filters separating branded from non-branded queries, with branded labels in the interface. For you, this means AI visibility splits by brand demand versus discovery, so measure non-branded citations as your true reach.

  • Claude adds connectors for Adobe, Blender, Autodesk Fusion, and creative tooling

    Anthropic has expanded the Claude Platform with new connectors covering tools that creative professionals already use: Blender, Autodesk Fusion, Adobe Creative Cloud apps, Ableton, Splice, Affinity by Canva, SketchUp, and Resolume. The connectors give Claude direct access to the design, audio and 3D-modelling environments where creative production happens, rather than only the file outputs of those workflows.

  • Claude adds enterprise-managed authentication for connectors

    Anthropic has introduced enterprise-managed authentication for Claude connectors, announced on its Claude for Business account. Admins can authorise and authenticate a connector once at the organisation level through their identity provider: starting with Okta: and users gain access the first time they log in, with access staying consistent across Claude chat, Claude Code, and Cowork.

  • Claude adds Microsoft 365 connectors across all plans

    Anthropic has launched Microsoft 365 connectors across all Claude plans, enabling users to connect Outlook, OneDrive, and SharePoint directly to Claude conversations, bringing email, documents, and files into context without manual attachment.

  • Claude adds Slack integration on Pro and Max plans

    Anthropic has launched Slack integration for Claude on Pro and Max plans. You can search workspace channels, prepare for meetings, and send messages without leaving the Claude interface.

  • Claude adds background computer use on macOS in beta

    Claude says computer use can now run in the background through Cowork and Claude Code while users work in another window.

  • Claude brings file creation, skills, and compaction to the Free plan

    Anthropic has made file creation and editing available on Claude's Free plan, alongside skills and compaction features previously limited to paid users. This enables free users to handle more complex tasks and to stay logged in for longer.

  • Claude can now create and edit Excel, PowerPoint, and PDF files directly in the browser.

    Anthropic has added native document creation to Claude, allowing users to generate Excel, PowerPoint, and PDF files through conversational prompts without leaving the browser. The feature removes the need for separate tools or API integrations for basic document automation.

  • Claude Code adds self-hosted environments in public beta

    Self-hosted environments let Claude Code sessions run on your own infrastructure, started from web, mobile, desktop, or a routine, next to internal services and security controls. For you, this means agentic coding can stay inside your network boundary, so evaluate it where compliance blocked cloud sessions.

  • Claude Code adds server previews and automated code review

    Anthropic has shipped significant updates to Claude Code on desktop, introducing server previews, automated code review, and PR monitoring with auto-fix capabilities. Server previews allow Claude to start development servers and preview running applications within the desktop interface, read console logs, catch errors, and iterate automatically.

  • Claude Code adds shareable Artifacts

    Anthropic has added Artifacts to Claude Code, announced on its Claude account. Artifacts are interactive pages built from a working session: a pull-request walkthrough or a living project dashboard, for example: shared with a team at a private link.

  • Claude Code can now run locally for free via Ollama

    Claude Code can now run for free using Ollama's Anthropic API compatibility. Developers can run Claude Code locally with open-source models while maintaining full agent and tool workflows.

  • Claude Code Channels enables remote control via Discord and Telegram

    Anthropic has launched Channels in Claude Code: a research preview feature that allows users to connect a persistent local Claude Code session to a Discord or Telegram bot, enabling remote task delegation from a phone or secondary device. Messages sent to the bot arrive as events in the running Claude session; Claude reads the request, executes work on the local machine, and replies back through the same channel.

  • Claude Code gains a Security plugin in beta

    The Claude Security plugin scans code for vulnerabilities before commits or across a codebase, from the terminal. It runs on the Claude inference you already pay for.

  • Claude Code gains auto mode for autonomous permission handling

    Anthropic has launched auto mode in Claude Code, available as a research preview on the Team plan, with Enterprise and API access rolling out over the following days. Auto mode allows Claude Code to handle file writes and bash command permissions autonomously, without requiring approval for each action or disabling safeguards entirely.

  • Claude Code gains Computer Use directly from the terminal

    Charly Wargnier has flagged that Claude Code now supports Computer Use directly from the terminal, enabling Claude to compile and launch macOS apps (SwiftUI or Electron), click through the UI to verify fixes, and complete tasks that previously required leaving the terminal. The feature is enabled via the /mcp command and requires macOS with a Pro or Max plan.

  • Claude Code integrates with Figma via MCP

    Claude Code can now push prototypes directly to Figma using the latest version of Figma's MCP server. The integration allows developers to take a working prototype built in Claude Code and send it to a Figma canvas, enabling exploration of different design directions without branching code or rebuilding from scratch.

  • Claude Code makes Auto mode the default from 14 August

    New Claude Code sessions on Pro, Max, and Team plans will run in Auto mode from 14 August, routing every action through a safety classifier instead of asking approval per command. For you, this means agentic coding gets faster with guardrails shifting to the classifier, so review the setting if your workflows need manual approval.

  • Claude Code now included with all Anthropic Team plan subscriptions

    Anthropic has announced that Claude Code is now included with every Team plan seat at no additional cost. Teams can use Claude Code to accelerate coding workflows including debugging, refactoring, and shipping features.

  • Claude Code Routines launch in research preview

    Anthropic has launched Routines in Claude Code in research preview across all paid plans. Routines allow a prompt, repository, and connectors to be configured once and run on a schedule, triggered by an API call, or fired by a GitHub webhook.

  • Claude Community Ambassadors programme opens applications

    Anthropic has launched the Claude Community Ambassadors programme, inviting builders to lead local Claude communities, host events, and partner with Anthropic to shape Claude's future. For you, this means opportunities to build local PPC and marketing AI communities with Anthropic support.

  • Claude connects to health data via Apple Health, Health Connect, and more

    Anthropic has launched health data integrations for Claude in beta. Four integrations are available: Apple Health (iOS), Health Connect (Android), HealthEx, and Function.

  • Claude Cowork adds a built-in browser

    Claude Cowork is gradually adding a browser inside the Claude Desktop app for macOS, Windows, and Linux, with availability on Pro, Max, Team, and enabled Enterprise plans. It can open sites, read pages, click, type, and complete forms in a side panel without relying on the user's normal browser; existing Claude in Chrome users keep the extension as the default unless they switch.