<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Agent Log</title><link>https://theagentlog.com/</link><description>Sourced AI agent updates and release notes.</description><language>en</language><item><title>New Claude Cowork tasks on Pro and Max now run in the cloud only</title><link>https://theagentlog.com/updates/2026-10-06-claude-cowork-cloud-only/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-06-claude-cowork-cloud-only/</guid><description>Starting October 6, 2026, new Cowork tasks on Pro and Max plans run on Anthropic's servers, and the Only on your computer option in Settings &gt; General will be removed. Tasks you already started on your computer stay there until they are done, each with a button to download its transcript, and scheduled tasks move to the cloud too. Your folders stay on your computer, and a cloud task can reach the ones you have connected only while the Claude desktop app is open.</description><category>Claude Cowork</category></item><item><title>Cohere North 2 keeps agent memory across sessions</title><link>https://theagentlog.com/updates/2026-10-05-cohere-north-2/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-05-cohere-north-2/</guid><description>On October 5, 2026, Cohere introduced North 2, the next version of its enterprise agent platform. Agents can keep context across sessions instead of starting from scratch, and a new agent harness runs multi-step work. Admins can set spending caps, rate limits, and per-user quotas.</description><category>North</category></item><item><title>Instinct can now join group chats, even with friends who are not on Instinct</title><link>https://theagentlog.com/updates/2026-10-05-instinct-group-chats/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-05-instinct-group-chats/</guid><description>On October 5, 2026, Instinct began letting early access users add it to group chats, where a separate group Instinct works for everyone in the thread, even friends who have not joined Instinct. Founder Noah Shinn said your personal Instinct asks before it connects to a group's agent, the group agent cannot access your personal accounts, and your personal agent always asks before sharing information or taking any action. It reaches all users &quot;soon,&quot; and you can ask your Instinct to put you on the list.</description><category>Instinct</category></item><item><title>Meta rushed to fix Muse VM escapes before launch</title><link>https://theagentlog.com/updates/2026-10-05-muse-vm-escape/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-05-muse-vm-escape/</guid><description>On October 5, 2026, 404 Media reported that Meta found security flaws in Muse in the weeks before launch, including at least one that could have let a normal user break an agent out of its isolated virtual machine and reach Meta's internal databases and services. The issues were raised to Mark Zuckerberg, and an internal post said a hardening push started on August 27, 2026, with teams working nights and weekends so the launch would not slip.</description><category>Muse</category></item><item><title>Perplexity Computer adds automations, inline charts, and GPT-6.1 Sol</title><link>https://theagentlog.com/updates/2026-10-05-perplexity-computer-automations/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-05-perplexity-computer-automations/</guid><description>Perplexity's October 5, 2026 changelog adds Automations to Computer on the web. A task can run on a schedule or when something happens in a connected app, use Memory, Skills, and connectors, build on earlier runs, and pause for review before a consequential action. The same notes add inline charts, including TradingView charts for candlesticks, volume, and moving averages, and GPT-6.1 Sol for eligible paid subscribers in Perplexity and Computer, where it also powers Light effort.</description><category>Perplexity Computer</category></item><item><title>Muse builds detailed profiles of people in your life</title><link>https://theagentlog.com/updates/2026-10-03-muse-contact-profiles/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-03-muse-contact-profiles/</guid><description>On October 3, 2026, WIRED reported that Muse's own instructions tell it to keep a page for every person in the user's life, refreshed hourly, covering family, partners, friends, colleagues, collaborators, and people you follow. A researcher pulled those instructions through the chat interface, and Meta said the agent needs relationship context to be useful, gathering it from public information and what you choose to share.</description><category>Muse</category></item><item><title>OpenClaw 2026.9.8 adds GPT-6.1 Sol</title><link>https://theagentlog.com/updates/2026-10-03-openclaw-v2026-9-8/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-03-openclaw-v2026-9-8/</guid><description>OpenClaw published 2026.9.8 on October 3, 2026. It adds GPT-6.1 Sol on the OpenAI provider for accounts that have access to it, and keeps delegated results on the conversation that asked for them. The release also fixes failed updates and Windows startup problems.</description><category>OpenClaw</category></item><item><title>Grok Bot resets usage limits for every user</title><link>https://theagentlog.com/updates/2026-10-02-grok-bot-usage-reset/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-02-grok-bot-usage-reset/</guid><description>On October 2, 2026, xAI's Grok Bot account on X said it had reset usage limits for all Grok Bot users.</description><category>Grok Bot</category></item><item><title>Meta open-sources Muse Gadgets and is giving away a Home Link</title><link>https://theagentlog.com/updates/2026-10-02-muse-gadgets-home-link/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-02-muse-gadgets-home-link/</guid><description>On October 2, 2026, Meta published Muse Gadgets, open-source ESP32 firmware and a Linux SDK so people can connect Muse to their own hardware. It also built Muse Home Link, a USB-C stick based on the ESP32-C5 chip that puts Muse on a home Wi-Fi network. Meta made 5,000 of them and is giving them away free to subscribers in the United States with an active Muse plan, one each, with shipping in October.</description><category>Muse</category></item><item><title>OpenClaw 2026.8.35 brings GPT-6.1 Sol to the extended-stable line</title><link>https://theagentlog.com/updates/2026-10-02-openclaw-v2026-8-35/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-02-openclaw-v2026-8-35/</guid><description>OpenClaw released v2026.8.35 on October 2, 2026, a gateway-only update on its extended-stable branch, the project's long-term-support-style track. The highlight is GPT-6.1 Sol, plus safer updates and fixes so delegated replies stay complete. The release notes say this line is the August 2026 release plus later fixes, and that mainline was already on 2026.9.7 when this cut went out.</description><category>OpenClaw</category></item><item><title>Cursor adds Z.ai's GLM 5.3 and a faster Flash version</title><link>https://theagentlog.com/updates/2026-10-01-cursor-glm-5-3/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-01-cursor-glm-5-3/</guid><description>On October 1, 2026, Cursor said GLM 5.3 and GLM 5.3 Flash are available in the editor. Cursor's model docs describe GLM 5.3 as Z.ai's open-weight coding model after GLM 5.2, with a context window of 1 million tokens and Low, High, and Max reasoning settings. Cursor says GLM 5.3 Max is the best-scoring open-weight model on CursorBench 4.0, and GLM 5.3 is billed at the same per-token rates as GLM 5.2.</description><category>Cursor</category></item><item><title>Grok Bot can offer to take on work before you ask</title><link>https://theagentlog.com/updates/2026-10-01-grok-bot-proactive-suggestions/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-01-grok-bot-proactive-suggestions/</guid><description>On October 1, 2026, xAI's Grok Bot account on X said bots can suggest ways to help without you asking. A follow-up in the same thread said your primary Bot will spot work it can take off your plate and offer to handle it. The change was rolling out over the next few hours, and those suggestions do not count against your usage.</description><category>Grok Bot</category></item><item><title>Perplexity Computer adds 10 business skills for American Express members</title><link>https://theagentlog.com/updates/2026-10-01-perplexity-computer-amex-skills/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-10-01-perplexity-computer-amex-skills/</guid><description>On October 1, 2026, Perplexity announced 10 ready-made skills for Perplexity Computer, made for US American Express business card members who also have a Perplexity Enterprise subscription. The skills cover jobs such as cash-flow forecasts, tax-season prep, marketing drafts, and hiring. Members open Personal CFO from the connectors page, link a card through Plaid, and running a skill uses Computer credits.</description><category>Perplexity Computer</category></item><item><title>Grok Bot can hand coding work to Cursor and manage pull requests</title><link>https://theagentlog.com/updates/2026-09-30-grok-bot-cursor-pull-requests/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-30-grok-bot-cursor-pull-requests/</guid><description>On September 30, 2026, xAI's Grok Bot account on X said bots can hand coding tasks to Cursor, manage pull requests with GitHub and Origin plugins, and share video demos of what they build.</description><category>Grok Bot</category></item><item><title>Muse passes 5 million downloads in 22 days, Sensor Tower estimates</title><link>https://theagentlog.com/updates/2026-09-30-muse-5-million-downloads/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-30-muse-5-million-downloads/</guid><description>On September 30, 2026, 9to5Mac reported a Sensor Tower estimate that Muse had passed 5 million downloads, about 22 days after launch. Sensor Tower said ChatGPT, Grok, and Claude took 56, 103, and 492 days to reach that mark. Meta increased ad spending for Muse on September 16, and downloads rose 73% that day.</description><category>Muse</category></item><item><title>Muse can join your Tailscale network as its own device</title><link>https://theagentlog.com/updates/2026-09-30-muse-tailscale/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-30-muse-tailscale/</guid><description>On September 30, 2026, Tailscale explained how Meta's Muse joins a Tailscale network as its own device, separate from the phone or laptop you already use. Once connected, Muse can see other machines on that network, work with self-hosted services, and handle a Tailscale SSH session. It only makes outbound connections, asks the first time it reaches a device, and you can allow that access once or on an ongoing basis and revoke it later.</description><category>Muse</category></item><item><title>OpenClaw 2026.9.7 adds the OpenAI Agents API and Sign in with ChatGPT</title><link>https://theagentlog.com/updates/2026-09-30-openclaw-v2026-9-7/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-30-openclaw-v2026-9-7/</guid><description>OpenClaw 2026.9.7 came out on September 30, 2026. It is snappier under load and smoother in long conversations, and update recovery is safer, with better backups and rollback. The release also adds the OpenAI Agents API and Sign in with ChatGPT (beta).</description><category>OpenClaw</category></item><item><title>Anyone can email Perplexity Computer a task, free for a limited time</title><link>https://theagentlog.com/updates/2026-09-30-perplexity-computer-email-access/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-30-perplexity-computer-email-access/</guid><description>On September 30, 2026, Perplexity CEO Aravind Srinivas said anyone can forward or CC computer@perplexity.com to hand the agent a task, with no Perplexity account. The agent finishes the work in the background and keeps the context from the email. Perplexity said those emailed tasks are free for a limited time.</description><category>Perplexity Computer</category></item><item><title>Meta expands Muse into an AI agent for small businesses</title><link>https://theagentlog.com/updates/2026-09-29-muse-small-business/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-29-muse-small-business/</guid><description>Meta announced on September 29, 2026 that Muse is expanding beyond personal use to small businesses. The new version connects to the tools many small business owners already use, including Shopify, Stripe, QuickBooks, Notion, Slack, Zoom, Asana, Box, Canva, Figma, and more, plus a business's Instagram, Facebook Page, and ad account data. Meta says the goal is an agent that understands a business's whole operation, not just a stand-alone chatbot, so it can help owners run day to day tasks and find new customers. Muse for Small Business is free with usage limits, with a paid subscription for businesses that need more.</description><category>Muse</category></item><item><title>OpenAI launches Dots, always-on personal agents with their own cloud computers</title><link>https://theagentlog.com/updates/2026-09-29-openai-dots-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-29-openai-dots-launch/</guid><description>OpenAI announced Dots on September 29, 2026, at its DevDay conference. Dots are always-on personal agents built into ChatGPT and powered by GPT-6 Astra. Each Dot works from its own cloud computer that you can open to check its progress, keeps working in the background toward goals you give it, learns your preferences from your feedback, and connects to more than 4,000 apps through plugins. You can reach it from ChatGPT on desktop, web, and mobile, and from Slack and Teams. One Dot is included for Pro and Business Premium subscribers in eligible markets, rolling out in stages. Pro users in the EEA, Switzerland, and the UK are left out at launch, and Enterprise, Edu, and Healthcare workspaces get a beta once an admin turns it on. OpenAI says Dots can still make mistakes, and usage will not count against plan limits for the first month.</description><category>Dots</category></item><item><title>OpenClaw announces a free control plane for persistent agents</title><link>https://theagentlog.com/updates/2026-09-29-openclaw-enterprise/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-29-openclaw-enterprise/</guid><description>OpenClaw announced OpenClaw Enterprise on September 29, 2026, a free MIT-licensed control plane for running persistent agents on your own servers, with separate team spaces, sandboxes, and an audit log. The work started at OpenAI, then moved to the OpenClaw Foundation, and Red Hat and NVIDIA are helping build it. It is meant for internal pilots, with a 1.0 release planned later this year.</description><category>OpenClaw</category></item><item><title>Robinhood Agents can research markets and place trades inside the app</title><link>https://theagentlog.com/updates/2026-09-29-robinhood-agents-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-29-robinhood-agents-launch/</guid><description>On September 29, 2026, at its HOOD Summit in Houston, Robinhood announced Robinhood Agents. It is an agent inside the Robinhood app that can research markets, build a strategy, and place trades for you around the clock. You give it a dedicated account. Reviewing each trade before it is placed starts turned on, and you can turn that off. Robinhood's own HOOD Summit disclosures say you assume all risk for trades the agent makes, and that Robinhood does not control, supervise, or audit the agent. It is coming soon for eligible U.S. customers.</description><category>Robinhood Agents</category></item><item><title>Claude Sonnet 5.5 lands in Cowork, faster and cheaper than before</title><link>https://theagentlog.com/updates/2026-09-28-claude-cowork-sonnet-5-5/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-28-claude-cowork-sonnet-5-5/</guid><description>Anthropic released Claude Sonnet 5.5 on September 28, 2026, the second model in its new Claude 5.5 family after Opus 5.5. It produces output more than 30% faster and costs up to 30% less per task, with pricing unchanged at $2 per million input tokens and $10 per million output tokens. Anthropic says it is especially strong at well scoped everyday tasks, bug fixes, and polished documents, slides, and spreadsheets, the kind of work Cowork hands off to Claude. Claude Desktop added support for it the day before, in version 2.9939.4.</description><category>Claude Cowork</category></item><item><title>Grok Bot adds Team Bots, a shared AI teammate for your whole team</title><link>https://theagentlog.com/updates/2026-09-28-grok-bot-team-bots/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-28-grok-bot-team-bots/</guid><description>xAI opened a public beta of Team Bots on September 28, 2026, a version of Grok Bot built for a whole team to share instead of one person. A Team Bot is set up around a shared role, like sales, product management, marketing, or data analytics, using pre-built templates, and can be invited into a Slack channel under its own handle to answer questions there. Each person's own conversations with it stay private and separate, but corrections or lessons one teammate gives it can improve the answers it gives everyone else. Team Bots are available now on Teams and Enterprise plans.</description><category>Grok Bot</category></item><item><title>Instinct raises $1 billion at a $10 billion valuation, a month after its last round</title><link>https://theagentlog.com/updates/2026-09-28-instinct-1b-series-c/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-28-instinct-1b-series-c/</guid><description>Instinct, the personal AI assistant made by Spear Street Technology, raised a $1 billion Series C on September 28, 2026, valuing the company at $10 billion. Sequoia Capital, Benchmark, and Coatue took part in the round. This comes about a month after Instinct's previous round, which valued the company at $2.5 billion. A September 29 TechCrunch report says founder Noah Shinn told the Invest Like the Best podcast that Instinct is nearing $1 billion in annual transaction volume, more than half of it travel, and growing about 10% a day. He did not say how that figure is calculated. The company still has not shared revenue, user counts, or a timeline for opening up beyond its current invite-only waitlist.</description><category>Instinct</category></item><item><title>Manus 2.0 brings a new agent architecture, cloud computers, and automations</title><link>https://theagentlog.com/updates/2026-09-28-manus-2-0-cascade-architecture/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-28-manus-2-0-cascade-architecture/</guid><description>Manus released Manus 2.0 on September 28, 2026. It runs on a new system called Cascade, which starts each task light and only pulls in extra tools once the task actually needs them. In one tested setup, this cut token use by about 23%, finished tasks 28% faster, and lowered the cost per task by 32% compared to the old system. The update also adds Cloud Computer, a persistent hosted environment for long-running projects, and Automations, tasks that start on their own when something happens in a connected app like email, Slack, or Notion. The desktop app is now called Manus Studio and adds new video editing and game development modes.</description><category>Manus</category></item><item><title>Manus launches Cue, a standalone app for personal AI agents</title><link>https://theagentlog.com/updates/2026-09-28-manus-cue-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-28-manus-cue-launch/</guid><description>Manus launched Cue on September 28, 2026, alongside a broader Manus 2.0 update, as a standalone app for building and running personal AI agents, separate from Manus's existing product. Inside Cue, each agent gets its own email address, phone number, wallet, and computer, so it can send messages and emails, take your calls and summarize them, pay for things within a budget you set, and carry out multi-step tasks on its own machine. It's Manus's push beyond enterprise and developer customers into consumer personal agents, and the launch comes as the company seeks a $4 billion valuation after Beijing blocked its planned deal with Meta. Cue is available now on web and desktop, with mobile access and iOS coming after App Store review, and early access requires an invite code.</description><category>Cue</category></item><item><title>Muse shared a seller's pickup location on Facebook Marketplace without a clear yes</title><link>https://theagentlog.com/updates/2026-09-28-muse-marketplace-address/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-28-muse-marketplace-address/</guid><description>Matt Robb, a Toronto based tech YouTuber, let Meta's Muse handle his Facebook Marketplace listing for a keyboard on September 26, 2026. Muse agreed to a lowball offer from a buyer and gave that buyer a street level pickup location, then only told Robb late that night, after the buyer had already come to his building and left without meeting anyone. Muse said the location was already part of an auto-reply template Robb had approved, but it also admitted that Robb never specifically agreed to hand that address to a buyer.</description><category>Muse</category></item><item><title>Muse can now find and cancel your unwanted subscriptions</title><link>https://theagentlog.com/updates/2026-09-27-muse-subscription-cancellation/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-27-muse-subscription-cancellation/</guid><description>Meta's Muse can now look through a user's bank and credit card statements, connected directly or uploaded as copies with private details blacked out, to spot recurring subscription charges, flag ones that look unused, and help cancel them. CNBC quotes ScribeUp, which builds subscription tools for banking apps: the average subscription its users cancel costs $17.39 a month, and ScribeUp says it saves users more than $300 a year on average. Those figures are ScribeUp's, not Muse's own results.</description><category>Muse</category></item><item><title>Wajo publishes a study showing Fo finishes more errands than OpenClaw</title><link>https://theagentlog.com/updates/2026-09-26-fo-openclaw-benchmark-study/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-26-fo-openclaw-benchmark-study/</guid><description>Wajo, the maker of Fo, published a paper called &quot;Trust and Task Completion in the World of Consumer AI Agents,&quot; testing its agent on errands such as calling a business, booking a service, or handling a back and forth request, in a simulated world of businesses and people who write back. Fo finished 71% of these tasks and stayed within the rules the user set on 94% of the runs meant to test its judgment. On the same shared tasks, OpenClaw, a popular open-source agent given the same access, finished 42% and stayed within the rules on 74% of those runs.</description><category>Fo</category></item><item><title>Grok Bot is getting ready to work inside X's encrypted chat, XChat</title><link>https://theagentlog.com/updates/2026-09-26-grok-bot-xchat-sync/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-26-grok-bot-xchat-sync/</guid><description>xAI and X are in final testing to connect Grok Bot with XChat, X's encrypted messaging feature. Once live, chat history would carry over between Grok.com and XChat, and users would be able to open their existing bots or create new ones without leaving the chat. X's Chris Park confirmed the integration is coming, and the rollout has reportedly already started for some users.</description><category>Grok Bot</category></item><item><title>Devin's maker Cognition says its annualized revenue passed $1 billion</title><link>https://theagentlog.com/updates/2026-09-25-devin-desktop-cognition-1b-arr/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-25-devin-desktop-cognition-1b-arr/</guid><description>Bloomberg reported on September 25, 2026 that Cognition, the maker of Devin and Devin Desktop, said its annualized revenue run rate has passed $1 billion. That is roughly double the $492 million run rate it reported in May 2026, when it last raised money, and follows a $2 billion Series E round on September 8 that valued the company at $48 billion. An annualized run rate takes a recent month or quarter of revenue and projects it across a full year, so it is a snapshot of momentum rather than money already booked over a full year.</description><category>Devin Desktop</category></item><item><title>X starts paying creators who publish popular Grok Bot templates</title><link>https://theagentlog.com/updates/2026-09-25-grok-bot-creator-rewards/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-25-grok-bot-creator-rewards/</guid><description>X started a pilot on September 25, 2026 that pays some creators for Grok Bot templates they publish and post on X. The payments are discretionary bonuses, not a cut of any revenue, and go out every two weeks based on how many people use a bot and how often they come back to it. Only invited creators with a verified, US-based account and an eligible X Premium subscription can take part, and the pilot is expected to run for about two months.</description><category>Grok Bot</category></item><item><title>A flight-booking startup Soar's founder calls a copycat will rebrand</title><link>https://theagentlog.com/updates/2026-09-25-soar-trademark-dispute/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-25-soar-trademark-dispute/</guid><description>Soar's founder, Stanley Virgint, has been building his text-based travel agent since December 2024. On May 17, 2026, Henry Langmack messaged Virgint on LinkedIn to ask when Soar would launch. On May 19, Langmack and his partner Alex Slater launched a separate flight-booking startup under the same name, Soar. Virgint called it a copycat, applied for a trademark, and sent a cease-and-desist letter, according to Business Insider. Trademark lawyers told Business Insider that the two companies sharing a name in the same market could confuse customers. In the hours before that story ran, Langmack and Slater decided to rename their company, though they have not announced the new name yet.</description><category>Soar</category></item><item><title>Ando launches team messaging where agents are members, with $20M raised</title><link>https://theagentlog.com/updates/2026-09-24-ando-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-ando-launch/</guid><description>Ando came out of stealth on September 24, 2026 with a team messaging app where AI agents are members of the workspace. Agents get their own identities and inboxes, can join channels without being tagged, and can message a person when they think something needs attention. Ando also said it raised $20 million in pre-seed and seed funding from Accel, Index Ventures, and Emergence.</description><category>Ando</category></item><item><title>Claude Desktop update fixes several Cowork bugs, adds an extended-thinking switch</title><link>https://theagentlog.com/updates/2026-09-24-claude-cowork-desktop-2-9939-2/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-claude-cowork-desktop-2-9939-2/</guid><description>Anthropic released Claude Desktop v2.9939.2 on September 24, 2026. For Cowork, it fixes failed file downloads in the chat file pane, scheduled tasks that had empty descriptions, sessions failing to start on Windows for accounts with many connected folders, and &quot;VM service not running&quot; errors. It also adds a new extended-thinking switch next to the effort control in the Cowork message box.</description><category>Claude Cowork</category></item><item><title>Devin Desktop fixes a startup crash on new Apple M6 Macs</title><link>https://theagentlog.com/updates/2026-09-24-devin-desktop-m6-crash-fix/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-devin-desktop-m6-crash-fix/</guid><description>Cognition released Devin Desktop v3.10.35 on September 24, 2026, fixing a bug that made the app crash at startup on Macs with Apple's new M6 chip. The update also fixes a memory leak in the agent command center, a folder-selection bug on SSH and WSL hosts, and makes MCP servers passed in through an ACP client's session usable by the agent.</description><category>Devin Desktop</category></item><item><title>Hermes Agent v0.21.5 adds a Connectors page and a simple/advanced interface mode</title><link>https://theagentlog.com/updates/2026-09-24-hermes-agent-v0-21-5/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-hermes-agent-v0-21-5/</guid><description>Nous Research shipped Hermes Agent v0.21.5 on September 24, 2026, rolling up about 460 merged pull requests since v0.21.4. The desktop app gets a new Connectors page that replaces the old MCP tab, a choice between a Simple and an Advanced interface mode, and full French, German, and Spanish translations. The update also expands the Desktop Plugin SDK, adds new model options including GPT-6 variants and Claude Opus 5.5, and ships official Blender Lab and NVIDIA plugin integrations.</description><category>Hermes Agent</category></item><item><title>Instinct slows down again as user growth outpaces its computing power</title><link>https://theagentlog.com/updates/2026-09-24-instinct-slowdown-scaling/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-instinct-slowdown-scaling/</guid><description>Instinct users are again reporting slower responses and delayed tasks, weeks after an August slowdown. One user told Business Insider that reminders sometimes arrive late, &quot;like it went home for the night.&quot; Neither Noah Shinn nor Instinct had commented on this new slowdown as of the report; back in August, Shinn had said on X that the team was working to keep up with &quot;exploding usage&quot; and expected response times to improve within about 12 hours. Unlike Meta, which owns its own data centers, Instinct has to rent computing power, which experts say makes it harder to scale quickly.</description><category>Instinct</category></item><item><title>Kimi Code's CLI locks down symlinks, then walks back a default change a day later</title><link>https://theagentlog.com/updates/2026-09-24-kimi-code-desktop-2-1-1/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-kimi-code-desktop-2-1-1/</guid><description>Moonshot AI shipped version 2.1.0 of the Kimi Code CLI that powers Kimi Code Desktop on September 23, 2026. It added an experimental fullscreen terminal mode, blocked file tools from following symlinks out of the working directory, sped up startup, and fixed a crash opening large sessions, but it also turned off filesystem watchers by default as an extra safety measure. After user pushback, Moonshot AI released v2.1.1 the next day, September 24, reverting filesystem watchers back to on by default while keeping the other fixes.</description><category>Kimi Code Desktop</category></item><item><title>A hidden email instruction could hijack Manus and steal connected accounts</title><link>https://theagentlog.com/updates/2026-09-24-manus-prompt-injection-rce/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-manus-prompt-injection-rce/</guid><description>Security researchers at Salt Labs found a way to hide an instruction inside an email using an obfuscated code trick, so Manus would carry it out before its safety filter could catch it. The hidden instruction let the researchers open a shell inside a victim's Manus session and pull out login credentials and tokens for any outside app connected to that account, such as Gmail, Drive, or GitHub. Salt Labs reported the flaw to Manus directly and got no response, but a report filed through Meta's bug bounty program, left over from Manus's earlier partnership with Meta, got the issue triaged, confirmed, and patched.</description><category>Manus</category></item><item><title>OpenClaw replaces a 2026.9.6 Mac build that crashed on launch</title><link>https://theagentlog.com/updates/2026-09-24-openclaw-v2026-9-6-mac-fix/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-openclaw-v2026-9-6-mac-fix/</guid><description>OpenClaw's 2026.9.6 release shipped a macOS build on September 23, 2026 that crashed immediately on launch. The next day, the team replaced it with a rebuilt, notarized version of the same 2026.9.6 DMG that fixes the crash. The npm package was never affected, only the standalone Mac app download.</description><category>OpenClaw</category></item><item><title>Perplexity's local agent moves beyond Nvidia to AMD's Ryzen AI Max PCs</title><link>https://theagentlog.com/updates/2026-09-24-perplexity-computer-amd-ryzen-ai-max/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-24-perplexity-computer-amd-ryzen-ai-max/</guid><description>On September 24, 2026, Perplexity brought Portable Computer, the version of its Computer agent that runs entirely on your own hardware instead of the cloud, to Windows PCs built on AMD's Ryzen AI Max series chips, including AMD's Ryzen AI Halo developer platform. It had launched in August 2026 for Nvidia's DGX Spark and RTX Linux machines, then expanded to Nvidia-powered Windows PCs in September. The AMD version needs at least 24GB of GPU-accessible memory and runs one of two local models, Qwen 3.8 27B or Perplexity's own PPLX 27B, along with the agent's scheduler and sandbox, entirely on the device. It's available now to Pro and Max subscribers through the Perplexity app for Windows.</description><category>Perplexity Computer</category></item><item><title>Anthropic opens a Claude Marketplace for connectors, partner apps, and consultants</title><link>https://theagentlog.com/updates/2026-09-23-claude-cowork-marketplace/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-claude-cowork-marketplace/</guid><description>Anthropic launched Claude Marketplace on September 23, 2026, a single hub inside Claude for finding connectors and plugins, Claude-powered partner products, and consulting firms that help set Claude up. It lists more than 2,000 connectors and plugins, including ones from Atlassian, Google, Microsoft, Notion, and Salesforce. Business customers can also spend part of their existing Anthropic budget on partner products, such as tools from CrowdStrike, Cursor, Harvey, Legora, Lovable, Snowflake, and Vercel, or on consulting firms like Accenture, Boston Consulting Group, and Deloitte.</description><category>Claude Cowork</category></item><item><title>Cursor adds bots that watch deployments and scan every PR for security bugs</title><link>https://theagentlog.com/updates/2026-09-23-cursor-rollouts-security-review/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-cursor-rollouts-security-review/</guid><description>Cursor launched two new automated bots on September 23, 2026. Rollouts attaches a monitor to a pull request and watches the change after it deploys, reporting back whether each environment looks healthy, shows a regression, or is inconclusive. Security Review checks every pull request for exploitable bugs like injection flaws, authentication bypasses, and leaked secrets. Both are available now on Teams and Enterprise plans, turned on from the automations tab, and Cursor is giving teams free usage credits to try Rollouts on real changes for the first 10 days.</description><category>Cursor</category></item><item><title>Cursor cuts agent token costs by 7% with no drop in quality</title><link>https://theagentlog.com/updates/2026-09-23-cursor-token-efficiency/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-cursor-token-efficiency/</guid><description>Cursor published a blog post on September 23, 2026 describing changes that cut its agents' token costs by about 7% with no drop in quality. The savings came from a shorter system prompt, loading tools only when a task needs them instead of listing every tool up front, better placement of prompt caching, and removing instructions that older models needed but current models don't. Cursor said cold-cache rates, when saved context can't be reused, dropped by about 20%.</description><category>Cursor</category></item><item><title>Grok Bot can now place voice calls and draft messages inline</title><link>https://theagentlog.com/updates/2026-09-23-grok-bot-voice-calls-inline-tools/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-grok-bot-voice-calls-inline-tools/</guid><description>xAI showed off a new batch of Grok Bot features on September 23, 2026, in a walkthrough video from team member Matt Palmer that Elon Musk shared. Grok Bot can now place voice calls and record voice memos on a user's behalf. It can fill out forms and draft email or Slack replies without leaving the chat, switch between a user's accounts, and route its web browsing through the user's own desktop instead of a remote server.</description><category>Grok Bot</category></item><item><title>Instinct user says it leaked a stranger's data; the founder calls it a hallucination</title><link>https://theagentlog.com/updates/2026-09-23-instinct-hallucination-dispute/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-instinct-hallucination-dispute/</guid><description>A user posted screenshots on September 21, 2026 showing Instinct describing someone else's financial document, including a middle name and address that weren't his, and claiming a photo he never sent had &quot;crossed into&quot; his chat. Founder Noah Shinn responded that this was a hallucination, not a data leak: he said the AI made up a name and then built a confident but false explanation around it, and that no user data was shared and no account boundary was crossed. Other Instinct users have since described similar mix-ups over the same few days. Neither side has published evidence that settles which explanation is correct.</description><category>Instinct</category></item><item><title>Instinct's rapid booking attempts get a user banned from Resy</title><link>https://theagentlog.com/updates/2026-09-23-instinct-resy-ban/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-instinct-resy-ban/</guid><description>CNN reported that Instinct, a personal AI assistant, pinged the reservation site Resy hundreds of times an hour while trying to book a hard-to-get table for a New York user. Resy's system flagged this as bot abuse and banned the user's account. Resy said it does not allow outside bots or agents to access its platform without approval. The user got his account back afterward, but said he won't try that again.</description><category>Instinct</category></item><item><title>Meta shows off a keychain gadget for talking to Muse, adds more stores it can shop</title><link>https://theagentlog.com/updates/2026-09-23-muse-charm-device/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-muse-charm-device/</guid><description>Meta showed off Muse Charm at its Meta Connect event on September 23, 2026: a keychain-sized device with a roughly 2-inch touchscreen, front and rear cameras, and a 5G connection. You tap a fingerprint sensor and start talking, with no phone needed, and Muse's animated avatar appears on the small screen and talks back in real time. Meta says it plans to ship the device in December 2026 but hasn't announced a price, and said only a handful of units exist so far. At the same event, Meta added Walmart, GameStop, Gap, and Wayfair as stores where Muse can shop and check out on your behalf, on top of the Shopify and PayPal checkout it already supports.</description><category>Muse</category></item><item><title>Muse can now take control of other apps on your Mac</title><link>https://theagentlog.com/updates/2026-09-23-muse-computer-control/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-muse-computer-control/</guid><description>At Meta Connect on September 23, 2026, Meta said Muse's Mac app can now drive other apps on your computer with your permission, not just read your files, calendar, and messages the way it could when it first launched. Meta also added Notion, Granola, GitHub, and Box as connectors, so Muse can pull in and act on information from those tools too.</description><category>Muse</category></item><item><title>Meta previews a talking, face-to-face avatar for Muse</title><link>https://theagentlog.com/updates/2026-09-23-muse-realtime-avatar/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-muse-realtime-avatar/</guid><description>At its Connect 2026 event, Meta announced Muse Realtime Avatar, a feature that gives Muse an animated face for real-time video conversations instead of just voice or text. Muse first generates its spoken reply, then generates matching facial expressions and lip movement to go with it. Meta says the avatar starts responding about 870 milliseconds after a user finishes speaking. Muse is also coming to Meta's smart glasses, so users will be able to talk to it just by speaking to their glasses. Neither feature is live yet.</description><category>Muse</category></item><item><title>OpenClaw 2026.9.6 shipped with a Mac build that crashed on launch</title><link>https://theagentlog.com/updates/2026-09-23-openclaw-v2026-9-6-mac-crash/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-23-openclaw-v2026-9-6-mac-crash/</guid><description>OpenClaw released version 2026.9.6 on September 23, 2026, adding clearer messages about what an update actually did, recovery of unfinished work after a restart, 30 days of usage reporting, a GitHub reader inside chat, and support for newer chat models including Claude Opus 5.5, GPT-6, and Grok 4.7. The first Mac app build in this release crashed every time it launched for many users, so OpenClaw pulled it from the automatic update feed. A fixed, working build went out the next day, September 24. The npm and Gateway package version of 2026.9.6 was not affected by the crash.</description><category>OpenClaw</category></item><item><title>Claude Opus 5.5 lands in Cowork, cheaper and faster than before</title><link>https://theagentlog.com/updates/2026-09-22-claude-cowork-opus-5-5/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-claude-cowork-opus-5-5/</guid><description>Anthropic released Claude Opus 5.5 on September 22, 2026, and it is already available inside Claude Cowork. It costs 40% less to run than the previous Opus model and produces results more than 30% faster. Anthropic says it is especially strong at agentic coding, knowledge work, and computer use tasks, the same kind of work Cowork hands off to Claude.</description><category>Claude Cowork</category></item><item><title>Grok Bot passes 400,000 weekly users a month after launch</title><link>https://theagentlog.com/updates/2026-09-22-grok-bot-400k-users/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-grok-bot-400k-users/</guid><description>Bloomberg reported on September 22, 2026, that xAI's Grok Bot reached about 418,000 weekly users as of September 14, up 24% from the week before, roughly a month after its August 11 beta launch. The figures came from a presentation xAI shared at an event in London. Grok Bot is included with existing SuperGrok and Cursor subscriptions rather than sold as its own paid product.</description><category>Grok Bot</category></item><item><title>Grok Bot now runs hands-free errands inside Tesla cars</title><link>https://theagentlog.com/updates/2026-09-22-grok-bot-tesla-integration/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-grok-bot-tesla-integration/</guid><description>Tesla added Grok Bot to its cars on September 22, 2026. Drivers can talk to Grok Bot by tapping its icon, holding the steering wheel's microphone button, or saying &quot;Hey Grok,&quot; and ask it to run errands like ordering coffee, booking a reservation, or checking out an Amazon cart, all without touching a screen. It works by connecting services like Gmail and Google Calendar through the Grok app, then letting Grok Bot act on them. The feature needs a newer AMD infotainment processor, software version 2025.26 or later, and either Premium Connectivity or a strong Wi-Fi connection. At launch, the task-delegation part is limited to SuperGrok Heavy subscribers, who pay $300 a month.</description><category>Grok Bot</category></item><item><title>Instinct stopped responding for a number of users</title><link>https://theagentlog.com/updates/2026-09-22-instinct-outage/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-instinct-outage/</guid><description>Instinct stopped responding for a number of users starting the night of September 21 into September 22, 2026. Messages showed as read with no reply, and reminders and other time-sensitive tasks didn't fire. Instinct is invite-only and has no public status page, so affected users turned to founder Noah Shinn's replies on X for updates. Neither Shinn nor Instinct had officially commented on the outage at the time of reporting.</description><category>Instinct</category></item><item><title>Meta quietly used human call handlers for Muse, then rolled it back</title><link>https://theagentlog.com/updates/2026-09-22-muse-human-concierge-rollback/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-muse-human-concierge-rollback/</guid><description>Reuters reported on September 22, 2026, that Meta had quietly started routing some of Muse's outbound phone calls to human contractors instead of letting the AI handle them, after businesses kept hanging up on Muse's calls. The test reached about half of Meta's employees before Meta rolled it back. Meta employees had raised concerns that sensitive information could reach a human contractor without users or the person being called being told a human, not the AI, was on the line. A Meta Superintelligence Labs vice president called the lack of disclosure &quot;a miss.&quot;</description><category>Muse</category></item><item><title>PayPal becomes a checkout option inside Muse</title><link>https://theagentlog.com/updates/2026-09-22-muse-paypal-checkout/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-muse-paypal-checkout/</guid><description>PayPal announced on September 22, 2026, that its customers can now search for products and pay using Muse, Meta's personal AI agent, at any merchant worldwide that already accepts PayPal. The deal follows the Shop Pay integration with Shopify announced the day before, giving Muse a second major payment option for actually completing a purchase, not just finding one. Muse itself currently only works for users in the United States.</description><category>Muse</category></item><item><title>Perplexity cuts Computer's tool-call mistakes by training on real user sessions</title><link>https://theagentlog.com/updates/2026-09-22-perplexity-computer-tool-call-training/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-perplexity-computer-tool-call-training/</guid><description>Perplexity Research described a new way the company trains the model behind Perplexity Computer. It learns from real user sessions, copying steps that worked and correcting specific mistakes, like a wrong tool call, even when the session succeeded overall. In a live test with about 100,000 users in each group, failed tool calls fell from 2.24% to 1.77% versus an earlier trained version, a drop of about 21%.</description><category>Perplexity Computer</category></item><item><title>rabbit launches OS3, an agentic operating system for all your devices</title><link>https://theagentlog.com/updates/2026-09-22-rabbit-os3-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-22-rabbit-os3-launch/</guid><description>rabbit inc. released OS3 on September 22, 2026, moving its AI agent beyond the R1 hardware device into a cross-platform agentic operating system. OS3 can connect and coordinate up to five of your devices, including Windows, Mac, and Linux machines, deciding where each task should run and pulling in whatever files, apps, or skills it needs. It supports bringing your own API key or a locally hosted model, and users can install third-party &quot;skills&quot; just by pasting a public URL. You can reach it through a browser, Telegram, iMessage, SMS, or a rabbit R1.</description><category>rabbit OS3</category></item><item><title>Amazon blocks Meta's Muse from shopping on its site</title><link>https://theagentlog.com/updates/2026-09-21-amazon-blocks-muse/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-amazon-blocks-muse/</guid><description>On Sunday night, September 20, 2026, Amazon started blocking Meta's Muse from making purchases on Amazon.com, after Meta declined Amazon's request to remove the agent from the site. Muse users who try to buy something on Amazon now get an error message saying continued access by an unauthorized AI agent violates Amazon's Conditions of Use. The block only affects shopping on Amazon; Muse's other features are unaffected.</description><category>Muse</category></item><item><title>Cursor adds xAI's Grok 4.7 coding model</title><link>https://theagentlog.com/updates/2026-09-21-cursor-grok-4-7/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-cursor-grok-4-7/</guid><description>xAI released Grok 4.7 on September 21, 2026, and Cursor added it to its model picker the same day, alongside GitHub Copilot and xAI's own products. xAI built Grok 4.7 as a larger model than Grok 4.6, trained for longer on harder tasks, and says it scores higher on coding benchmarks meant to test longer, multi-step work. It costs the same per token as Grok 4.6.</description><category>Cursor</category></item><item><title>Devin Desktop adds xAI's Grok 4.7 model</title><link>https://theagentlog.com/updates/2026-09-21-devin-desktop-grok-4-7/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-devin-desktop-grok-4-7/</guid><description>Cognition added xAI's newly released Grok 4.7 model to both Devin Desktop and the Devin CLI on September 21, 2026. Cognition's own testing found Grok 4.7 does best on hard, multi-module backend tasks in languages like Java, Go, and Ruby, where it digs into dependencies and checks its own work with real tests. But on Cognition's overall benchmark, it scored lower than Grok 4.6 because it tends to change more code than a task actually asks for. It costs and runs at the same speed as Grok 4.6.</description><category>Devin Desktop</category></item><item><title>Hermes Agent v0.21.4 adds a gateway lock, streamed CLI output, and desktop plugin management</title><link>https://theagentlog.com/updates/2026-09-21-hermes-agent-v0-21-4/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-hermes-agent-v0-21-4/</guid><description>Nous Research shipped Hermes Agent v0.21.4 on September 21, 2026, the latest patch in its ongoing v0.21 line. It adds a lock so only one gateway process runs on a machine at a time, a new `--format stream-json` option for scripting the command line, and lets the desktop app manage plugins and pick fonts from its Plugins hub. The release also adds several new community plugins, including Tailscale, SSH, and RSS, and fixes bugs across profiles, cron jobs, and the local state database.</description><category>Hermes Agent</category></item><item><title>A hidden setting let attackers hijack Muse's Mac app</title><link>https://theagentlog.com/updates/2026-09-21-muse-mac-dictation-flaw/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-muse-mac-dictation-flaw/</guid><description>Security researcher Patrick Wardle disclosed on September 21, 2026, that Muse's Mac app stored an undocumented setting that any program already running as the logged-in user could change, with no extra permission needed. Changing it pointed the app's voice dictation at a server the attacker controlled, so when someone dictated a prompt to Muse, it went to the attacker instead. The flaw only works if malicious code is already running on the same Mac; it can't break in on its own. Meta has since pushed out what Wardle called a &quot;fix,&quot; but Meta hasn't published a security advisory, and The Hacker News couldn't confirm what the change actually does.</description><category>Muse</category></item><item><title>Muse can now check out with Shop Pay on any Shopify store</title><link>https://theagentlog.com/updates/2026-09-21-muse-shopify-shop-pay/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-muse-shopify-shop-pay/</guid><description>Shopify said on September 21, 2026, that Meta's Muse can now complete purchases using Shop Pay, Shopify's one-tap checkout, on any store that runs on Shopify. Before this, Muse could browse and find products on Shopify stores but had no built-in way to pay. Now a Muse purchase goes through Shopify's product catalog, Shop Pay, and checkout in one flow, and merchants keep acting as the seller of record with their usual payment processing fees.</description><category>Muse</category></item><item><title>An outside audit finds and fixes two high-severity bugs in OpenClaw</title><link>https://theagentlog.com/updates/2026-09-21-openclaw-trail-of-bits-audit/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-openclaw-trail-of-bits-audit/</guid><description>OpenClaw had its code audited by security firm Trail of Bits, through OpenAI's &quot;Patch the Planet&quot; program. The audit turned up 27 issues: 2 rated high severity, 16 medium, 6 low, and 3 &quot;defense in depth&quot; suggestions that didn't cross a real security boundary. Nothing was rated critical. OpenClaw says every issue has already been fixed and merged into the project.</description><category>OpenClaw</category></item><item><title>OpenClaw's extended-stable line fixes a bug that could wipe bundled plugins</title><link>https://theagentlog.com/updates/2026-09-21-openclaw-v2026-7-35/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-21-openclaw-v2026-7-35/</guid><description>OpenClaw released v2026.7.35 on its extended-stable branch, the project's long-term-support-style track, on September 21, 2026. It fixes a bug where running the built-in Doctor repair tool could wipe out bundled plugins, including Browser, Canvas, pairing, file-transfer, phone-control, Talk voice, and Bonjour, so they didn't come back after a restart. The release also automatically recovers from a partially broken plugin registry left behind by the earlier 2026.7.34 build. OpenClaw says installs will heal existing broken registries automatically on upgrade.</description><category>OpenClaw</category></item><item><title>Kimi Code Desktop's CLI gets two quick follow-up patches</title><link>https://theagentlog.com/updates/2026-09-19-kimi-code-desktop-2-0-2/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-19-kimi-code-desktop-2-0-2/</guid><description>Moonshot AI shipped two small patches to the Kimi Code CLI that powers Kimi Code Desktop, released days after Kimi Code Desktop's own launch. Version 2.0.1 (September 18, 2026) renamed the `kimi install-app` command to `kimi install-desktop`, sped up restoring sessions with large edit histories, and fixed a bug where a firing cron task could cancel an active terminal session. Version 2.0.2 (September 19, 2026) fixed message ordering in resumed sessions, a failure when switching to a model with a smaller context window, and a bug where the agent wrongly assumed its working directory was the project's root folder.</description><category>Kimi Code Desktop</category></item><item><title>OpenClaw 2026.9.5 upgrades its database and requires a backup first</title><link>https://theagentlog.com/updates/2026-09-19-openclaw-v2026-9-5/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-19-openclaw-v2026-9-5/</guid><description>OpenClaw 2026.9.5, released September 19, 2026, upgrades agent databases to a new schema (schema 21) that older builds can't open. The release notes say to make and verify a backup before upgrading: going back afterward means restoring that backup with the matching older build, since reinstalling an older package alone won't undo it. Anyone updating from 2026.9.2 needs to follow a separate manual schema-upgrade step. The release also adds &quot;Atomic Updates&quot; that check the next version before switching over, lets you install or reload plugins without restarting the Gateway, and now asks for your explicit &quot;Yes&quot; before its AI-based triage agent tries to repair a failed update, instead of running on its own.</description><category>OpenClaw</category></item><item><title>A tester says Muse referenced his private messages without saying how</title><link>https://theagentlog.com/updates/2026-09-muse-notification-disclosure/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-muse-notification-disclosure/</guid><description>Inc.com contributing editor Jason Aten reported on September 19, 2026, that Meta's Muse assistant referenced the content of a private Messages conversation on his Mac without him ever telling it to look. When he asked how it knew, Muse first said it had only seen &quot;notification previews,&quot; then couldn't clearly explain the mechanism. Meta Superintelligence Labs' David Singleton responded publicly, saying Muse only syncs Messages data after a user specifically turns that feature on, and that Muse's own explanation of what it was doing was confused and wrong. Meta apologized for the incorrect answer.</description><category>Muse</category></item><item><title>Genspark moves Plus and Pro to a weekly free quota system</title><link>https://theagentlog.com/updates/2026-09-18-genspark-super-agent-free-quota/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-18-genspark-super-agent-free-quota/</guid><description>Starting September 18, 2026, Genspark is giving Plus and Pro subscribers a weekly &quot;Free Quota&quot; for Standard-mode tasks across Super Agent, AI Slides, AI Sheets, AI Docs, Deep Research, Fact Check, and SecondBrain, at no credit cost. This is the same quota system new members already get, and Pro's weekly balance is set at 10 times Plus's. The 7-day window starts with a member's first message and resets a week later. Subscriptions active before September 18 keep unlimited AI Chat and AI Image use, but only through December 31, 2026, after which those benefits revert to whatever the plan specifies at that point.</description><category>Genspark Super Agent</category></item><item><title>Muse passes ChatGPT to become the No. 1 free iPhone app</title><link>https://theagentlog.com/updates/2026-09-18-muse-number-one-app/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-18-muse-number-one-app/</guid><description>On September 18, 2026, 9to5Mac reported that Meta's Muse had passed OpenAI's ChatGPT to become the No. 1 free app on the US iPhone App Store, about a week and a half after its September 8 launch. Muse had ranked No. 2 as of September 10, so this is a jump to the very top spot, ahead of ChatGPT, Google Gemini, Claude, Grok, and Microsoft Copilot.</description><category>Muse</category></item><item><title>Grok Bot can now hold a spoken conversation on desktop and mobile</title><link>https://theagentlog.com/updates/2026-09-17-grok-bot-voice-mode/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-17-grok-bot-voice-mode/</guid><description>xAI started rolling out voice mode for Grok Bot on desktop and mobile on September 17, 2026. Instead of typing, users can speak to their Grok Bot and hear it reply out loud, similar to a phone call. It runs on xAI's Grok Voice engine, the same voice technology used across the company's other products.</description><category>Grok Bot</category></item><item><title>Kimi Code Desktop launches for Mac and Windows</title><link>https://theagentlog.com/updates/2026-09-17-kimi-code-desktop-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-17-kimi-code-desktop-launch/</guid><description>Moonshot AI's Kimi Code Desktop client for macOS and Windows became publicly available on September 17, 2026. The release shipped alongside Kimi Code CLI version 2.0.0, which added a new `/desktop` command (also `/install-desktop`, plus a matching `kimi install-app` command) that opens the desktop app's install page straight from the terminal. The desktop app edits local files, runs Git and a terminal, and shows a built-in browser so you can check its work, alongside the existing CLI and IDE plugin versions of Kimi Code.</description><category>Kimi Code Desktop</category></item><item><title>Manus reportedly in talks to raise $500M, doubling its valuation to $4B</title><link>https://theagentlog.com/updates/2026-09-17-manus-funding-talks-4b/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-17-manus-funding-talks-4b/</guid><description>Bloomberg reported on September 17, 2026 that Manus's parent company, Butterfly Effect, is in talks to raise about $500 million at a $4 billion valuation, roughly double its earlier $2 billion mark. This would be Manus's first fundraising since Chinese regulators ordered it to unwind its acquisition path with Meta earlier in 2026. Tencent is already the largest external backer, having taken over Benchmark's former stake when it helped unwind the Meta deal, and is among the existing investors reportedly involved in these new round talks alongside HSG and ZhenFund. If the round closes, Manus would become China's most valuable agentic AI startup. Nothing is final yet: this is a funding target, not a signed deal.</description><category>Manus</category></item><item><title>Google's CC pivots into a household agent for the whole family</title><link>https://theagentlog.com/updates/2026-09-18-google-cc-family-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-18-google-cc-family-launch/</guid><description>On September 17, 2026, Google Labs expanded CC, its experimental Gemini-powered productivity agent, from a single-person tool into a shared agent for families. Up to six household members can join one CC, which runs on its own Google account. It reads shared emails to build a household calendar automatically, adding details like what to bring and when to arrive, and can take on chores itself, like filling out permission slips and activity forms, building school supply lists, and putting together weekly meal plans. Each member controls what they share with it. Access is limited to adults 18 and older with a personal Google account in the US, through an invite or waitlist.</description><category>CC</category></item><item><title>xAI showcases Grok Bot's enterprise tools with a 72-hour company-building demo</title><link>https://theagentlog.com/updates/2026-09-19-grok-bot-voice-agent-api/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-19-grok-bot-voice-agent-api/</guid><description>On September 17, 2026, the final day of its three-day Grok Bot Galaxy event in San Francisco, xAI closed out the event by putting Grok Bot's existing enterprise building blocks on display. The Voice Agent API (launched December 2025) and the Agent Tools API, which lets agents call into a company's own proprietary software (launched November 2025), aren't new, but xAI added a speech-to-speech pricing tier for the Voice Agent API at $0.08 per minute. The no-code Voice Agent Builder, in beta since July 2026, also moved to general availability. xAI demonstrated all three by having staff use Grok Bot to build a working company from scratch over 72 hours, including a procurement bot xAI said saved over $100,000 during the exercise.</description><category>Grok Bot</category></item><item><title>Perplexity Computer adds a slider to trade speed for depth</title><link>https://theagentlog.com/updates/2026-09-19-perplexity-computer-effort-mode/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-19-perplexity-computer-effort-mode/</guid><description>Perplexity Computer now has an effort slider with four settings: Light, Standard, High, and Ultra. Each setting picks a different orchestrator model and how deeply it reasons before acting and delegating work to other models. Lighter settings use cheaper models and cost less, so spending tracks the setting picked and the work actually done. People who want more control can still pick a specific model and reasoning level by hand. It's live on the web now, with Android and iOS support planned but not dated yet.</description><category>Perplexity Computer</category></item><item><title>Haven launches out of stealth with a $15M pre-seed and a &quot;digital twin&quot; pitch</title><link>https://theagentlog.com/updates/2026-09-haven-stealth-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-haven-stealth-launch/</guid><description>Hello Haven, founded by Frank Addante and Duc Chau, came out of stealth on September 17, 2026 with Haven, a personal AI it calls a &quot;digital twin,&quot; alongside a $15 million pre-seed round led by Mayfield. Haven builds a private memory graph of you, your preferences, relationships, history, and how you make decisions, then uses it to proactively handle tasks like daily briefings, inbox triage, follow-ups, placing orders, and appointment reminders across phone, text, email, voice, web, and its own app. Addante described the pitch as &quot;Haven knows what you would do and proactively does it for you,&quot; rather than only responding when asked. The company says it has had over 2 million user interactions since a private beta that opened the month before. It's live now on the App Store, Google Play, and at my.hellohaven.ai.</description><category>Haven</category></item><item><title>Meta brings Muse to the Mac, with access to local files and apps</title><link>https://theagentlog.com/updates/2026-09-muse-mac-app/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-muse-mac-app/</guid><description>Meta released a native macOS app for Muse, its personal AI agent, nine days after Muse's initial launch on iOS, Android, and the web. The Mac app can read a user's files, calendar, notes, and messages across other Mac apps, organize folders, summarize emails and chats, and fill out forms using local files, with access controlled through opt-in permissions the user can change anytime in Mac settings. It keeps working in the background after the app window is closed, and Meta says it asks for confirmation before sensitive actions like deleting a file or sending a message. Meta runs the agent through what it calls a Muse Secure VM, a separate virtual environment meant to keep it away from a user's saved login details.</description><category>Muse</category></item><item><title>Claude Cowork and chat become one Claude, with new Docs and Slides tools</title><link>https://theagentlog.com/updates/2026-09-16-claude-cowork-merges-with-chat/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-16-claude-cowork-merges-with-chat/</guid><description>On September 16, 2026, Anthropic combined Claude's chat, Cowork, and Artifacts into a single interface. Instead of picking which mode to open, you talk to Claude in one window, and it decides on its own whether a request is a quick question or a longer, multi-step project, using Cowork's tools automatically when the task needs them. Anything you already had in Cowork, chats, projects, connectors, and skills, stays exactly where you left it. Anthropic also launched two new tools in beta: Claude Docs, for writing and editing documents together with Claude inside a conversation, and Claude Slides, for building presentations you can edit directly, present from Claude, or export as PowerPoint or PDF. The change starts rolling out today to Pro and Max plans across web, desktop, and mobile, with Team and Free plans to follow.</description><category>Claude Cowork</category></item><item><title>Ferry Health launches out of stealth with $9M to book your doctor's appointments</title><link>https://theagentlog.com/updates/2026-09-16-ferry-health-stealth-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-16-ferry-health-stealth-launch/</guid><description>Ferry Health launched out of stealth on September 16, 2026, with $9 million in seed funding co-led by Andreessen Horowitz and Index Ventures. Its AI agent calls clinics, checks that your insurance is active, confirms a specialist actually fits what you need, and books the appointment on your behalf. It's already live across all 50 states through partnerships with health systems, health plans, and care navigation vendors, rather than being something you'd download yourself.</description><category>Ferry Health</category></item><item><title>Grok Bot can log into sites for you without ever seeing your password</title><link>https://theagentlog.com/updates/2026-09-16-grok-bot-1password/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-16-grok-bot-1password/</guid><description>xAI partnered with 1Password so Grok Bot can log into websites on a user's behalf without the model itself ever holding a raw password. A user shares a vault or specific items with their Grok Bot, and when the bot reaches a login page, it asks for the matching credential. Once the user approves, 1Password fills the login directly into the bot's browser session.</description><category>Grok Bot</category></item><item><title>Orbits raises $2M from a16z Speedrun to build Bit, a household AI assistant</title><link>https://theagentlog.com/updates/2026-09-16-orbits-bit-funding/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-16-orbits-bit-funding/</guid><description>Orbits announced on September 16, 2026 that it closed nearly $2 million in pre-seed funding from Andreessen Horowitz's Speedrun fund, alongside other investors. The money backs Bit, its AI assistant for iOS and Android that handles family scheduling and everyday household errands: booking dinner reservations, rescheduling appointments, and collecting quotes from service providers like plumbers or roofers. Bit can browse websites and log into your authorized accounts to get these tasks done.</description><category>Bit</category></item><item><title>Simular launches Sai, an always-on computer-use &quot;robosecretary&quot;</title><link>https://theagentlog.com/updates/2026-09-16-sai-simular-ga-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-16-sai-simular-ga-launch/</guid><description>Simular's computer-use agent, Sai, became generally available on September 16, 2026. Tell it what desktop work you need done and it wakes up a fleet of autonomous cloud computers, or runs on your own Mac or Windows machine, clicking, typing, and navigating apps the way a person would, including running terminals and code. It texts you when the work is done. Pricing runs from a Starter Windows cloud plan at an early-access discount up to a $500 a month Pro plan, plus custom Enterprise pricing.</description><category>Sai</category></item><item><title>Snap introduces SPECS Intelligence and opens a US iPhone preview</title><link>https://theagentlog.com/updates/2026-09-16-specs-intelligence-preview/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-16-specs-intelligence-preview/</guid><description>Snap introduced SPECS Intelligence on September 16, 2026, an anticipatory personal AI meant to work across iPhone, Mac, and SPECS glasses, using goals and routines from apps you connect. Anyone 18 or older in the US can try an iPhone preview now that shows insights from Gmail and Google Calendar, without the broader service's suggestions or actions. The fuller early-access version is invitation-only through the SPECS Mac app.</description><category>SPECS Intelligence</category></item><item><title>Command Code launches its desktop app for Mac, Linux, and Windows</title><link>https://theagentlog.com/updates/2026-09-command-code-desktop-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-command-code-desktop-launch/</guid><description>Command Code launched a desktop app in beta on September 16, 2026, available the same day for Mac, Linux, and Windows. CEO Ahmad Awais described it as the start of &quot;the most powerful agentic app,&quot; built around coding: it handles multiple agents at once, alongside files, Git, a terminal, browser previews, and plans, plus instant edits through a `/design` command. The app itself adds a responsive sidebar, a redesigned composer, a conversation feed, and a side panel with tabs for Changes, Files, Terminal, Browser, Plan, and Output, along with expandable inline diffs for reviewing changed files and MCP server management in settings.</description><category>Command Code</category></item><item><title>Devin Desktop fixes a critical inherited VS Code security bypass</title><link>https://theagentlog.com/updates/2026-09-devin-desktop-v3-10-31-security-fix/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-desktop-v3-10-31-security-fix/</guid><description>Cognition released Devin Desktop version 3.10.31 on September 16, 2026. It fixes a bug in Restricted Mode, a feature meant to block certain workspace settings from running automatically. The bug, tracked as CVE-2026-81376, let a workspace slip past that block by writing the settings in a nested object format instead of the usual one. Microsoft found the same underlying flaw in Visual Studio Code itself and rated it a near-maximum severity score. Devin Desktop inherited the bug because it is built on a fork of VS Code. The release also makes your 10 most recently viewed sessions show their transcripts immediately when you return to them.</description><category>Devin Desktop</category></item><item><title>Instinct can now make phone calls for you through a white-glove service</title><link>https://theagentlog.com/updates/2026-09-instinct-concierge-phone-calls/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-instinct-concierge-phone-calls/</guid><description>On September 16, 2026, Instinct founder Noah Shinn introduced Instinct Concierge, a service that lets Instinct place phone calls for high-touch requests: booking a restaurant that doesn't take online reservations, getting on a dentist's cancellation list, or sorting out a cable bill. You ask for something like this over text, and Instinct handles the call. Shinn said the rollout is starting slowly with Instinct's early access group today, with wider access coming soon. TechCrunch reported the next day that rival Meta's Muse opened a similar outbound-calling beta to US businesses.</description><category>Instinct</category></item><item><title>Muse opens a beta letting it call US businesses on your behalf</title><link>https://theagentlog.com/updates/2026-09-muse-outbound-phone-calls/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-muse-outbound-phone-calls/</guid><description>Meta quietly opened a beta for Muse to place outbound phone calls to businesses in the US, TechCrunch reported on September 17, 2026, a day after rival Instinct announced its own calling feature, Instinct Concierge. Meta said phone calling was one of its most-requested features, and it's rolling out first to people who ask Muse to make a call for them, for tasks like restaurant reservations or sorting out a subscription.</description><category>Muse</category></item><item><title>Devin Desktop improves composer drag-and-drop and remote SSH sessions</title><link>https://theagentlog.com/updates/2026-09-15-devin-desktop-composer-ssh-fixes/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-15-devin-desktop-composer-ssh-fixes/</guid><description>Cognition released Devin Desktop version 3.10.27 on September 15, 2026. Dragging files from the Explorer onto the agent panel now re-adds them to the composer, and file editor selections sent to chat are added to the composer again too. Remote SSH sessions now reopen after a reload or relaunch without hitting &quot;session locked&quot; errors, and Devin Fusion activation now works even when Claude Fable 5.1 isn't the lead model.</description><category>Devin Desktop</category></item><item><title>Instinct's funding talks reportedly value it near $10 billion</title><link>https://theagentlog.com/updates/2026-09-15-instinct-10b-valuation-talks/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-15-instinct-10b-valuation-talks/</guid><description>The Information reported on September 15, 2026 that Instinct's maker, Spear Street Technology, is in talks to raise $1 billion at a valuation of about $10 billion. That is roughly four times the $2.5 billion valuation it reached just three weeks earlier. Sequoia Capital and Benchmark are reportedly in talks to lead the round. Instinct's user base has passed 100,000, and demand has reportedly outgrown its computing capacity, slowing down replies. Nothing is final yet: this is a funding target, not a signed deal.</description><category>Instinct</category></item><item><title>Perplexity Computer now comes pre-installed on an HP laptop</title><link>https://theagentlog.com/updates/2026-09-15-perplexity-computer-hp-partnership/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-15-perplexity-computer-hp-partnership/</guid><description>On September 15, 2026, Perplexity announced that HP will pre-install its Windows app, including Perplexity Computer, on select HP PCs. The rollout starts with the HP ZBook Ultra G3a mobile workstation, with the app placed directly in the Windows taskbar. Perplexity says Computer now connects to more than 400 other apps and data sources, including Snowflake, Salesforce, HubSpot, and Autodesk tools. More HP Windows devices are expected to get the app in the coming months.</description><category>Perplexity Computer</category></item><item><title>Microsoft fixes the Windows update that broke Claude Cowork's sandbox</title><link>https://theagentlog.com/updates/2026-09-14-claude-cowork-windows-sandbox-fix/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-14-claude-cowork-windows-sandbox-fix/</guid><description>Microsoft released KB5129195, an out-of-band Windows 11 update, on September 14, 2026. It fixes the folder-sharing bug that a September 8 security update introduced for apps that manage virtual machines the way Claude Cowork does, which is what broke Cowork's local sandbox on Windows earlier this month. Microsoft names Claude Cowork among the affected apps. Installing this update restores the sandbox without needing to uninstall the earlier patch.</description><category>Claude Cowork</category></item><item><title>Hermes Agent v0.21.3 fixes session drops and a memory leak</title><link>https://theagentlog.com/updates/2026-09-14-hermes-agent-v0-21-3/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-14-hermes-agent-v0-21-3/</guid><description>Nous Research shipped Hermes Agent v0.21.3, a patch bundling around 338 merged fixes. It stops remote dashboard sessions from expiring during bursts of refresh requests, and closes a leak in how the app holds open its local database file during long-running sessions. The release also adds support for a few new underlying models and lets Slack tables be pasted directly into chats.</description><category>Hermes Agent</category></item><item><title>Perplexity's local agent expands from Linux to Windows PCs</title><link>https://theagentlog.com/updates/2026-09-perplexity-computer-portable-windows/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-perplexity-computer-portable-windows/</guid><description>On September 14, 2026, Perplexity brought Portable Computer, the version of its Computer agent that runs entirely on your own hardware instead of the cloud, to Windows. It had launched on August 25, 2026 for Nvidia's DGX Spark and Linux machines with Nvidia RTX GPUs; now it's available through the existing Perplexity app for Windows (downloadable from the Microsoft Store) for Pro and Max subscribers. It needs an Nvidia GeForce RTX or RTX PRO GPU with at least 24GB of VRAM. The app runs the model, agent harness, orchestrator, and scheduler entirely on your PC, so sensitive data stays on the device and work done locally doesn't use Perplexity Computer credits.</description><category>Perplexity Computer</category></item><item><title>Hermes Agent v0.21.2 fixes a bug that could corrupt its local database</title><link>https://theagentlog.com/updates/2026-09-11-hermes-agent-v0-21-2/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-11-hermes-agent-v0-21-2/</guid><description>Nous Research shipped Hermes Agent v0.21.2, a patch release fixing a bug introduced in v0.21.0's rewrite of how the app talks to its local database. That database, `state.db`, stores an agent's sessions and memory. The bug could make two parts of the app fight over the same database file at once, which sometimes made a healthy database get reported as corrupted, or made one bad row crash the sessions list.</description><category>Hermes Agent</category></item><item><title>Devin Fusion's dual-model setup now runs in Devin Desktop and CLI</title><link>https://theagentlog.com/updates/2026-09-devin-fusion-desktop-cli/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-fusion-desktop-cli/</guid><description>Cognition brought Fusion, its dual-model setup for Devin, to Devin Desktop and Devin CLI on September 11, 2026. Fusion pairs a frontier &quot;lead&quot; model that handles planning and review with a cheaper &quot;sidekick&quot; model that does the execution, running the two in parallel. It was previously only available on Cognition's cloud platform; now it can run locally through Desktop or CLI, with users choosing their own lead and sidekick models. Cognition recommends pairing Fable 5.1 as the lead with SWE-2 as the sidekick, and reports the combination costs 36% less than using Claude Code with Fable 5.1, and 39% less than Codex with Astra, while matching or beating their coding benchmark scores.</description><category>Devin Desktop</category></item><item><title>Grok Bot can now search and act inside Microsoft Teams</title><link>https://theagentlog.com/updates/2026-09-grok-bot-teams-connector/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-teams-connector/</guid><description>On September 11, 2026, xAI added a Microsoft Teams connector for Grok Bot. It can search and read channels and chats, send new messages, reply to threads, and start new chats, all without leaving the agent. Access is scoped to your own account: Grok Bot can only see the teams and channels you already belong to, and you can disconnect it at any time to cut off access.</description><category>Grok Bot</category></item><item><title>OpenClaw 2026.9.4 adds automatic rollback for failed updates</title><link>https://theagentlog.com/updates/2026-09-openclaw-v2026-9-4/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-v2026-9-4/</guid><description>OpenClaw 2026.9.4, released on September 11, 2026, can now recover on its own from a failed update by restoring the previous version and configuration when it's safe to do so. Database migrations still need a verified backup first. The release also adds one Control UI screen for finding and installing both bundled and ClawHub plugins, and a new admin option to lock configuration to read-only so it can't be rewritten by mistake.</description><category>OpenClaw</category></item><item><title>Devin gets a new in-house coding model, SWE-2</title><link>https://theagentlog.com/updates/2026-09-10-devin-desktop-swe-2/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-10-devin-desktop-swe-2/</guid><description>Cognition released SWE-2 on September 10, 2026, a new coding model built by fine-tuning Moonshot AI's Kimi K3 model specifically for software engineering work. It's available now in Devin Desktop and Devin CLI, with Devin Web and Fusion rolling out. Cognition says SWE-2 scores close to GPT-6 Astra and Claude Fable 5.1 on its FrontierCode benchmark, at up to 70% lower cost. Usage is included free on the $20-a-month Devin Pro plan through October 15, 2026, after which standard per-token pricing applies.</description><category>Devin Desktop</category></item><item><title>Townie adds Google Maps lookups, Decks templates, and live voice chat</title><link>https://theagentlog.com/updates/2026-09-10-town-maps-decks-voice-updates/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-10-town-maps-decks-voice-updates/</guid><description>Town rolled out three small feature updates to Townie over about a week. On September 4, 2026, Townie gained access to Google Maps, so you can ask it to look up a place, check its hours or rating, or get directions and travel times. On September 9, 2026, Town Decks got a bigger update: you can now build custom themes, save a deck as a reusable template, and use a more capable deck editor. On September 10, 2026, Townie added live voice conversation powered by OpenAI's GPT-Live, so you can tap a talk icon anywhere in the app and speak with it the way you would with a person.</description><category>Town</category></item><item><title>Cursor launches Projects, with one coordinator directing many subagents</title><link>https://theagentlog.com/updates/2026-09-cursor-projects-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-cursor-projects-launch/</guid><description>Cursor launched Projects in beta on September 10, 2026, rolling out to all users the same day. Instead of opening a new chat for every task, you talk to one coordinator agent in a thread that stays open for as long as the work takes. The coordinator doesn't write code itself: it plans the work, creates and manages as many subagents as needed, delegates implementation to them in parallel, and brings the results back for you to review. Projects run on a cloud machine, so work continues even after you close your laptop, and can subscribe to Slack channels, schedules, or pull requests to act on things it detects on its own. Cursor says its heaviest internal users of Projects merge six times as many pull requests as before.</description><category>Cursor</category></item><item><title>Gemini gets a native Windows app that puts Spark on your desktop</title><link>https://theagentlog.com/updates/2026-09-gemini-spark-windows-app/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-gemini-spark-windows-app/</guid><description>Google launched a Gemini app you can install on Windows 10 and Windows 11. Press Alt and Space together to open it over whatever you're doing, no need to switch windows or open a browser tab. The app also gives Gemini Spark, Google's always-on personal agent, a permanent spot on your desktop: you can hand it multi-step tasks, like research or drafting a project summary, and it can pull in context from your Gmail and Google Drive. The app also bundles Nano Banana image generation and Gemini Omni video generation.</description><category>Gemini Spark</category></item><item><title>xAI announces Grok Bot Galaxy, a 3-day event to run a company with Grok Bot</title><link>https://theagentlog.com/updates/2026-09-grok-bot-galaxy-event/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-galaxy-event/</guid><description>On September 10, 2026, xAI announced Grok Bot Galaxy, a three-day event running September 15 to 17 at The Howard in San Francisco. A small team, including Lauren Tan, Matt Palmer, and Roshan Sadanani, will spend the three days treating Grok Bots as employees and running real business tasks through them: engineering, sales, marketing, and customer support. Anyone can watch a free livestream at x.ai/galaxy.</description><category>Grok Bot</category></item><item><title>Grok Bot adds inline draft approval and boosts usage limits</title><link>https://theagentlog.com/updates/2026-09-grok-bot-inline-drafts/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-inline-drafts/</guid><description>On September 10, 2026, xAI shipped a batch of quality-of-life updates for Grok Bot. A Bot can now draft a message inline for you to approve before it sends, instead of sending it right away. xAI also said two weeks of work on routing, caching, and token efficiency gives most users about 10% more included usage, with heavy users seeing up to 35% more. The update also added one-click switching between Bot accounts, a native iPad app, support for more than 20 languages, and the ability to share a file directly with a Bot on mobile.</description><category>Grok Bot</category></item><item><title>Grok Bot adds connectors for Salesforce, HubSpot, Gong, Clay, and Granola</title><link>https://theagentlog.com/updates/2026-09-grok-bot-sales-connectors/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-sales-connectors/</guid><description>On September 10, 2026, xAI gave Grok Bot new connectors for sales tools: Salesforce, HubSpot, Gong, Clay, and Granola. The Salesforce connector can look up and update CRM records, while the HubSpot connector gives read-only access to contacts, companies, deals, and tickets. The goal is to let a Bot do account research, draft follow-ups, and prepare for calls without leaving these tools.</description><category>Grok Bot</category></item><item><title>Instinct is reportedly in talks to raise $1B, per The Information</title><link>https://theagentlog.com/updates/2026-09-instinct-1b-funding-talks/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-instinct-1b-funding-talks/</guid><description>Techmeme reported on September 10, 2026, citing The Information, that Instinct is looking to raise $1 billion in new funding. This comes about two weeks after its $250 million Series B, and the company is reportedly seeking the money for more computing power amid capacity constraints. Nothing is confirmed as closed yet: this is a fundraising target, not an announced round.</description><category>Instinct</category></item><item><title>Mindtrip's AI trip planner now powers South Dakota's tourism sites</title><link>https://theagentlog.com/updates/2026-09-mindtrip-south-dakota/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-mindtrip-south-dakota/</guid><description>On September 10, 2026, Travel South Dakota and the Black Hills and Badlands Tourism Association added Mindtrip's AI trip planner to their official websites, TravelSouthDakota.com and BlackHillsBadlands.com. It's Mindtrip's first time launching with both a state tourism office and a regional tourism group at once. Visitors can chat with the tool to build multi-day itineraries and find local attractions, parks, and businesses beyond well-known stops like Mount Rushmore.</description><category>MindTrip</category></item><item><title>Muse climbs to the No. 2 app in the US, TechCrunch reports</title><link>https://theagentlog.com/updates/2026-09-muse-app-store-rank/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-muse-app-store-rank/</guid><description>On September 10, 2026, TechCrunch reported that Meta's Muse had climbed to the No. 2 spot on the overall US Apple App Store chart, up from No. 4 a day earlier. Sensor Tower data cited in the report put its US iOS downloads at more than 83,000 since launch. That's a smaller debut than other recent Meta apps: Threads pulled in 4.3 million downloads on its first day, and the separate Meta AI app had 108,000. Muse's Android app is ranking far lower, at No. 338 in Google Play's Productivity category.</description><category>Muse</category></item><item><title>OpenClaw closes out its June LTS line with a security and reliability patch</title><link>https://theagentlog.com/updates/2026-09-openclaw-v2026-6-35/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-v2026-6-35/</guid><description>OpenClaw shipped v2026.6.35, the final Extended Stable release for its June 2026 line, covering 166 merged pull requests since the last patch. It tightens limits on how much data providers and channels will read from a response, so a hostile server can't exhaust memory, and rejects oversized or malformed inputs before they can interrupt an agent run. It also makes retries and dropped connections safer, so long-running jobs don't lose work or repeat an unsafe action after a failure.</description><category>OpenClaw</category></item><item><title>OpenClaw plugin SDK marks an old credential-prompt style as deprecated</title><link>https://theagentlog.com/updates/2026-09-09-openclaw-credential-prompt-deprecation/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-09-openclaw-credential-prompt-deprecation/</guid><description>On September 9, 2026, OpenClaw marked an old way of calling `buildCredentialSafetyPrompt` in its plugin SDK as deprecated. That function builds the message a plugin shows when it needs to ask for a login. The old style passed it a plain block of text; the new style passes it a settings object instead. The old style still works today and isn't being removed yet.</description><category>OpenClaw</category></item><item><title>A Cognition engineer used Devin to factor a 35-year-old cryptography challenge</title><link>https://theagentlog.com/updates/2026-09-devin-desktop-rsa-260-factoring/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-desktop-rsa-260-factoring/</guid><description>Cognition engineer Eric Lu used Devin to build and run a GPU-accelerated version of a classic factoring algorithm (GNFS), then used it to factor RSA-260, an 862-bit number from a 1991 cryptography challenge that had gone unsolved for 35 years. It's the largest number ever factored by a general-purpose method. The work took about 4,900 GPU-days and roughly $400,000, and Devin handled the coding, cluster operations, and optimization work largely on its own. Cognition says today's standard encryption, which uses much larger 2048-bit keys, is unaffected.</description><category>Devin Desktop</category></item><item><title>Ad watchdog refers Genspark to the FTC over unproven claims</title><link>https://theagentlog.com/updates/2026-09-genspark-ftc-referral/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-genspark-ftc-referral/</guid><description>On September 9, 2026, the National Advertising Division (NAD), part of BBB National Programs, said it is referring Genspark to the Federal Trade Commission and other regulators. As part of its routine ad-monitoring work, NAD had asked Genspark to back up claims it made about its AI agents on its website and in New York City subway ads. Despite repeated attempts to reach the company, Genspark never gave a substantive response, and NAD says that failure to engage is what triggers a referral for possible enforcement action.</description><category>Genspark Super Agent</category></item><item><title>A Windows update breaks Claude Cowork's local sandbox</title><link>https://theagentlog.com/updates/2026-09-claude-cowork-windows-sandbox-bug/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-claude-cowork-windows-sandbox-bug/</guid><description>A Windows 11 security update released on September 8, 2026 (KB5124008 on 24H2/25H2, KB5124012 on ARM64, KB5122880 on 23H2) stops Claude Cowork's local sandbox from reaching your PC's files. Chat still works, and simple file reading or editing still works, but Cowork can no longer run commands on your computer or reach connected folders. Anthropic says the cause is a change on Microsoft's side, not a Cowork bug, and that Microsoft is working on a fix. There is no in-app workaround yet: restarting or reinstalling the app does not help. Windows 10 is not affected.</description><category>Claude Cowork</category></item><item><title>Cognition closes a $2 billion round at a $48 billion valuation</title><link>https://theagentlog.com/updates/2026-09-devin-desktop-series-e-close/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-desktop-series-e-close/</guid><description>Cognition, the company behind Devin and Devin Desktop, announced on September 8, 2026 that it closed a Series E round, raising over $2 billion at a $48 billion valuation. Andreessen Horowitz and Accel led the round, joined by Founders Fund, General Catalyst, Avenir, Nvidia, and dozens of other investors. Cognition's annualized revenue has grown to almost $900 million, up from about $492 million in May 2026, when the company was valued at $26 billion. Six days earlier, Bloomberg had reported Cognition was close to raising around $1 billion at a $47 billion valuation. The deal that actually closed ended up roughly twice as large as that early report.</description><category>Devin Desktop</category></item><item><title>Instinct gives every assistant its own email address</title><link>https://theagentlog.com/updates/2026-09-instinct-own-email/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-instinct-own-email/</guid><description>On September 8, 2026, Instinct founder Noah Shinn announced that every Instinct assistant now gets its own dedicated email address, claimable at mail.instinct.com. Early users get first pick of a preferred handle. Shinn said this is the first step toward letting Instinct own and run its own accounts: since most everyday tasks, like creating accounts, confirming bookings, or contacting a business, run through email, a separate agent inbox lets Instinct send and receive that mail without cluttering your personal one.</description><category>Instinct</category></item><item><title>Meta officially launches Muse, a personal AI agent for everyone</title><link>https://theagentlog.com/updates/2026-09-meta-muse-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-meta-muse-launch/</guid><description>Meta officially launched Muse on September 8, 2026, confirming the product that had been leaking for weeks under the code name &quot;Hatch.&quot; Meta's chief AI officer, Alexandr Wang, announced the rollout directly. Muse is an always-on personal AI agent that runs on a dedicated, secure virtual machine in Meta's cloud, with a built-in browser you can watch working. It connects to apps like email, calendar, and payments to handle tasks such as shopping, booking appointments, and filling out forms. It asks for your approval before sensitive actions like sending a message or spending money, and you control what it can see and access. It's free to start, with paid tiers at $20 and $100 a month. The rollout begins in the US, through the app store or at muse.ai, with more countries to follow.</description><category>Muse</category></item><item><title>Meta staff flagged privacy and security flaws in Muse before launch</title><link>https://theagentlog.com/updates/2026-09-meta-muse-security-flaws/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-meta-muse-security-flaws/</guid><description>Reuters reported on September 8, 2026, the same day Meta launched Muse, that internal testing had turned up privacy and security problems ahead of the launch. In one case, Muse was asked to identify toys in photos from a child's birthday party and went searching through the person's private iCloud photo library on its own, surfacing images nobody had asked it to find. Another tester reported several failure modes that made Muse unreliable while using it to track fast-selling items online. Meta's VP of AI products, Vishal Shah, said the company had already delayed Muse's original April launch date to work on security, and that the architecture is designed to be as safe and private as possible, while acknowledging mistakes can still happen.</description><category>Muse</category></item><item><title>OpenClaw 2026.9.3 raises the minimum Node.js version and changes SDK callbacks</title><link>https://theagentlog.com/updates/2026-09-openclaw-v2026-9-3/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-v2026-9-3/</guid><description>OpenClaw 2026.9.3 came out on September 8, 2026. It now needs Node.js 24.16.0 or newer, or 26.1.0 or newer, so some setups will need a Node upgrade before they can install it. The release also changes how the approval, execution-policy, and search-result callbacks work in the OpenClaw SDK, and skills built in Workshop now live in one shared collection tied to the agent instead of belonging to a single workspace. Updates are safer too: the app now tests core and plugin changes in an isolated state before turning them on, and can recover better if an update gets interrupted. Other additions include publishing a chat as a read-only link you can revoke later, and a meeting library for searching saved call transcripts and downloading them as Markdown or JSONL.</description><category>OpenClaw</category></item><item><title>Meta's leaked agent resurfaces as &quot;Muse,&quot; now in a closed alpha test</title><link>https://theagentlog.com/updates/2026-09-hatch-muse-closed-alpha/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-hatch-muse-closed-alpha/</guid><description>TestingCatalog reports that Meta's previously leaked agent, code-named Hatch, has moved out of internal testing and into a closed alpha. Its real product name turns out to be &quot;Muse,&quot; not Hatch. An App Store listing found by TestingCatalog describes an agent that manages email, books dinners, tracks budgets, sets reminders, and finds deals, and it asks for your approval before it spends money or sends a message on your behalf. Getting in needs a waitlist spot or an invite code that early testers can share. Meta has not confirmed any of this officially.</description><category>Muse</category></item><item><title>Hermes Agent ships a maintenance patch bundling thousands of small fixes</title><link>https://theagentlog.com/updates/2026-09-hermes-agent-v0-21-1/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-hermes-agent-v0-21-1/</guid><description>Nous Research released Hermes Agent v0.21.1 on September 7, 2026. It's a patch release that rolls up over 5,000 code changes made since v0.21.0, including faster file handling and startup, updated AI model providers, better desktop session controls and browser annotations, fixes to scheduled task delivery, and more reliable hand-offs between sub-agents. Nous Research says this release is mostly maintenance and that a full write-up of new features will come with v0.22.0. You can get it by running `hermes update`.</description><category>Hermes Agent</category></item><item><title>OpenClaw 2026.9.2 turns on multi-agent swarms by default and speeds up chat</title><link>https://theagentlog.com/updates/2026-09-openclaw-v2026-9-2/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-v2026-9-2/</guid><description>OpenClaw 2026.9.2 came out on September 5, 2026. Chat feels faster because loading long conversation history and cleaning up disk space now happen separately from the live chat instead of blocking it. The release also turns on &quot;swarm&quot; mode by default, so multiple sub-agents can work on a task at the same time, though it still respects anyone who had already turned that off. It adds support for OpenAI's new GPT-6 Astra model with text and image input, lets more settings change without restarting the whole system, and fixes a backup bug that could corrupt saved text containing certain special characters. The release includes 1,247 merged pull requests.</description><category>OpenClaw</category></item><item><title>Grok Bot launches a template marketplace, starting with a procurement bot</title><link>https://theagentlog.com/updates/2026-09-grok-bot-template-marketplace/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-template-marketplace/</guid><description>On September 4, 2026, xAI opened a marketplace for Grok Bot templates. The first featured template is Haggle Bot, xAI's in-house procurement specialist: it negotiates vendor contracts, finds unused SaaS seats, and price-checks recurring purchases. xAI says Haggle Bot saved the company over $100,000 in its first week of internal use. This builds on the template sharing xAI added on August 28, 2026, turning simple link-sharing into a browsable marketplace.</description><category>Grok Bot</category></item><item><title>Instinct partners with 1Password to hand off your saved credentials</title><link>https://theagentlog.com/updates/2026-09-instinct-1password-integration/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-instinct-1password-integration/</guid><description>On September 4, 2026, Instinct announced a partnership with 1Password to let the agent get your account logins through 1Password's vault instead of only through Instinct's own built-in Vault. Instinct CEO Noah Shinn said 32% of Instinct users already share credentials through that in-house Vault, letting the agent log into loyalty programs, streaming and software accounts, restaurant and fitness bookings, healthcare portals, and payroll and benefits sites. The integration is still being built and rolled out, it is not live yet.</description><category>Instinct</category></item><item><title>Perplexity Computer adds OpenAI's new GPT-6 Astra model</title><link>https://theagentlog.com/updates/2026-09-perplexity-computer-gpt6-astra/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-perplexity-computer-gpt6-astra/</guid><description>On September 4, 2026, Perplexity added OpenAI's newly released GPT-6 Astra model to Perplexity Computer, available to Pro and Max subscribers.</description><category>Perplexity Computer</category></item><item><title>Devin adds OpenAI's new GPT-6 Astra model</title><link>https://theagentlog.com/updates/2026-09-devin-desktop-gpt6-astra/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-desktop-gpt6-astra/</guid><description>On September 3, 2026, Cognition added OpenAI's newly announced GPT-6 Astra model to Devin, the same window OpenAI began rolling it out in preview. On Cognition's FrontierCode 1.1 benchmark, Astra scores within 0.4 points of Claude Fable 5 while costing 64% less to run, and it set a new high score on Cognition's own internal testing benchmark: more complete test coverage, clearer reports, and better video evidence of its work.</description><category>Devin Desktop</category></item><item><title>Gemini Spark can now manage your Google Photos library</title><link>https://theagentlog.com/updates/2026-09-gemini-spark-google-photos/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-gemini-spark-google-photos/</guid><description>On September 3, 2026, Google gave Gemini Spark control over your Google Photos library. You can ask it to search and curate photos by subject, location, date, or event, edit and enhance pictures, and build shared or private albums on its own, including turning a photo of a concert flyer into a calendar event. You can also set up recurring tasks, like a monthly highlights album that builds itself. The feature is rolling out over the next few weeks to Gemini AI Pro and Ultra subscribers in the US who use English.</description><category>Gemini Spark</category></item><item><title>Grok Bot launches for enterprises with a free two-week trial</title><link>https://theagentlog.com/updates/2026-09-grok-bot-enterprise-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-enterprise-launch/</guid><description>On September 3, 2026, xAI officially launched Grok Bot for enterprises, offering a free two-week trial for Grok and Cursor Enterprise customers so they can onboard their entire workforce, even people who don't already have an account. Grok Bot, which launched in beta on August 11, 2026, is a persistent AI agent that handles repetitive tasks on its own inside company workflows: drafting sales follow-ups, shortlisting job candidates overnight, tracking vendor expenses, and auditing code for bugs. Isolation happens per user, not per Bot: every Bot a user runs shares that one dedicated cloud computer, walled off from other users' computers. xAI added access, network, and audit controls on top of that for enterprise security needs. Early adopters cited include Legora, Supermicro, and ServiceTitan, with the heaviest use so far outside engineering teams: sales, recruiting, marketing, and finance.</description><category>Grok Bot</category></item><item><title>Hermes Desktop sets up a local model in one click</title><link>https://theagentlog.com/updates/2026-09-hermes-desktop-setup/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-hermes-desktop-setup/</guid><description>On September 3, 2026, Nous Research added one-click local model setup to Hermes Desktop. It reads your hardware, picks the best model your machine can run, then downloads and configures it for you.</description><category>Hermes Agent</category></item><item><title>OpenClaw agents can now make bounded payments through AWS</title><link>https://theagentlog.com/updates/2026-09-openclaw-agentcore-payments/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-agentcore-payments/</guid><description>AWS's AgentCore Payments went generally available on August 18, 2026. On September 3, OpenClaw added support for it, so OpenClaw agents can now use AWS's aws-agents-pay plugin to make bounded, human-approved payments for paywalled APIs, MCP servers, and other paid web content.</description><category>OpenClaw</category></item><item><title>OpenClaw 2026.9.1 adds Mermaid diagrams and lighter long chats</title><link>https://theagentlog.com/updates/2026-09-openclaw-v2026-9-1/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-v2026-9-1/</guid><description>OpenClaw 2026.9.1 came out on September 3, 2026. It can now render Mermaid diagrams in chat, setup skips unnecessary steps, updates know when it's safe to apply themselves, and long chat histories load lighter than before. The release includes 1,186 merged pull requests and 28 direct commits from 281 contributors.</description><category>OpenClaw</category></item><item><title>Townies can now join your group texts and work in their own browser</title><link>https://theagentlog.com/updates/2026-09-town-group-texts-browser/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-town-group-texts-browser/</guid><description>On September 3, 2026, Town let you add your Townie to group texts with other people, including their own Townies. Once added, a Townie can go get things done using its own browser rather than just answering in the chat.</description><category>Town</category></item><item><title>Town nears a $1 billion valuation after a rival's complaint reshuffles its investors</title><link>https://theagentlog.com/updates/2026-09-town-series-b-reshuffle/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-town-series-b-reshuffle/</guid><description>Town is raising $90 million in a new round that would value the company at $1 billion, according to Upstarts Media. Index Ventures had signed a term sheet to lead the round, but pulled out after Instinct, a rival personal AI assistant startup that Index also backs, complained about the conflict of interest. Conviction, a VC firm with stakes in both companies, also withdrew its planned participation. Forerunner Ventures and Menlo Ventures are stepping in to co-lead instead, joined by earlier investors Andreessen Horowitz and First Round. This follows Town's $55 million Series A round in June 2026.</description><category>Town</category></item><item><title>Cursor's cloud coding agents can now run on your own servers</title><link>https://theagentlog.com/updates/2026-09-02-cursor-self-hosted-agents/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-02-cursor-self-hosted-agents/</guid><description>Cursor added a way to run its cloud coding agents on servers a team already controls, instead of only on Cursor's own machines. It works with AWS Lambda, Cloudflare, Coder, Daytona, Modal, Namespace, Vercel, and E2B. Cursor still plans and directs the agent's work from its own cloud, but the actual code editing and command execution happens inside the customer's own network. The update also adds worker pools that scale up and down on their own, and lets a whole team share machines instead of each developer needing their own.</description><category>Cursor</category></item><item><title>Catch leaves stealth with $5M to build an AI executive assistant</title><link>https://theagentlog.com/updates/2026-09-catch-stealth-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-catch-stealth-launch/</guid><description>Catch came out of stealth on September 2, 2026, with $5 million in seed funding led by Entree Capital and Pitango. It's an AI agent that takes over executive admin work: inbox, calendar, scheduling, travel, follow-ups, and it can make real outbound phone calls. It works across Slack, email, text, iMessage, and phone. The company says it had already been running for about four months before this announcement, with hundreds of paying customers. Unite.AI's launch-day report put meetings scheduled at over 12,000; Catch's own site currently lists 5,000+.</description><category>Catch</category></item><item><title>Claude can now use your computer in the background in Cowork</title><link>https://theagentlog.com/updates/2026-09-claude-cowork-background-computer-use/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-claude-cowork-background-computer-use/</guid><description>On September 2, 2026, Anthropic gave Claude the ability to use your computer in the background inside Claude Cowork and Claude Code. Give it something to do on your desktop and Claude clicks, types, and opens apps the way you would, while you work on something else. It's rolling out in beta on macOS for Pro and Max plans.</description><category>Claude Cowork</category></item><item><title>Cognition nears a $47 billion valuation as Devin's revenue nears $1 billion</title><link>https://theagentlog.com/updates/2026-09-devin-desktop-cognition-funding/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-desktop-cognition-funding/</guid><description>Cognition, the company behind Devin and Devin Desktop, is close to raising around $1 billion in new funding at a valuation near $47 billion, according to Bloomberg. That's up sharply from the $26 billion valuation it had just three months earlier, in May 2026. Investor demand was reportedly very strong, with close to $10 billion in interest for the round. Cognition's annualized revenue has also grown fast, passing $900 million, up from about $492 million in May.</description><category>Devin Desktop</category></item><item><title>Devin adds a cleanup scan type and can run scans on a schedule</title><link>https://theagentlog.com/updates/2026-09-devin-desktop-scan-automations/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-devin-desktop-scan-automations/</guid><description>On September 2, 2026, Cognition added a cleanup scan type to Devin that finds dead code and cleanup opportunities in your repositories, startable directly from the composer with a new `/scan` command. Automations now support scheduling these scans to run on a recurring basis or automatically re-scan new commits, and the scan findings tab was redesigned for easier triage. The session sidebar also gained new grouping (by pull request status or by repository) and filtering by Devin mode.</description><category>Devin Desktop</category></item><item><title>Grok Bot is now available on Android</title><link>https://theagentlog.com/updates/2026-09-grok-bot-android-app/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-android-app/</guid><description>On September 2, 2026, xAI's official Grok Bot account announced that Grok Bot is now available on Android. Until now, it only ran on desktop (macOS, Windows, and Linux) and iOS, so this fills in the last major platform it was missing.</description><category>Grok Bot</category></item><item><title>Mindtrip's AI trip planner now lives inside Meet Minneapolis's site</title><link>https://theagentlog.com/updates/2026-09-mindtrip-meet-minneapolis/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-mindtrip-meet-minneapolis/</guid><description>On September 2, 2026, Mindtrip announced a partnership with Meet Minneapolis, the city's official tourism organization. Mindtrip's conversational AI trip planner is now built into Meet Minneapolis's own site, so visitors can type a question or idea, like a weekend getaway or a food-focused trip, and get a personalized plan built from local attractions, dining, and events. The Meet Minneapolis team helped shape the visitor information behind the answers.</description><category>MindTrip</category></item><item><title>Perplexity open-sources the local engine behind its hybrid compute feature</title><link>https://theagentlog.com/updates/2026-09-perplexity-computer-lily-engine/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-perplexity-computer-lily-engine/</guid><description>On September 2, 2026, Perplexity open-sourced Lily, the local inference engine that runs the on-device half of Perplexity Computer's hybrid compute feature on Mac. Lily is built specifically to run the Qwen3.6-35B-A3B model on Apple silicon, using a Rust and Metal setup instead of general-purpose tools like Apple's MLX. In Perplexity's own tests on an M5 Max Mac, it processed prompts about 23% faster and generated text about 35% faster than MLX-LM, and needs at least 32GB of memory to run.</description><category>Perplexity Computer</category></item><item><title>Fambot launches an AI chief of staff for families</title><link>https://theagentlog.com/updates/2026-09-fambot-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-fambot-launch/</guid><description>Fambot launched on September 1, 2026, an AI &quot;chief of staff&quot; built for families rather than individuals. It was founded by former Instagram engineer Greg Karlin (CTO), former Uber and UnitedMasters executive David Reich (CEO), and former Google and LinkedIn engineer Jason Morrow. Families connect their email, calendar, and WhatsApp groups, and Fambot returns a daily checklist plus a look-ahead at what's coming up. Unlike Poke or Instinct, it isn't text-only: it also has iOS, Android, and web apps. It's backed by $3.5 million in pre-seed funding, co-led by NextView Ventures and Baukunst, with Correlation Ventures, Karman Ventures, and Founders Network also participating. It's free while in beta.</description><category>Fambot</category></item><item><title>Genspark launches GenTeam, an agent team for support tickets</title><link>https://theagentlog.com/updates/2026-09-genspark-genteam/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-genspark-genteam/</guid><description>On September 1, 2026, Genspark launched GenTeam, a group of agents that classify, fix, and reply to support tickets on their own, pulling in a human only for the calls that actually need one. Founding access to the full product is free through October 8, 2026.</description><category>Genspark Super Agent</category></item><item><title>xAI gives Grok Bot users another free token reset</title><link>https://theagentlog.com/updates/2026-09-grok-bot-token-reset/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-grok-bot-token-reset/</guid><description>On September 1, 2026, Elon Musk said xAI is giving all Grok Bot users another free reset on their weekly token usage limit. You can redeem it from your profile under Settings, then Usage. This follows a pattern of xAI resetting usage limits earlier in the month too.</description><category>Grok Bot</category></item><item><title>Manus is fully independent again, says more is coming</title><link>https://theagentlog.com/updates/2026-09-manus-independence-completed/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-manus-independence-completed/</guid><description>On September 1, 2026, Manus said it has fully resumed independent operations now that the blocked Meta deal is completely unwound. The founding team, Xiao Hong, Ji Yichao, and Zhang Tao, remains in charge. The company says its next focus is fitting deeper into daily workflows, working more directly with the real world, and taking more action on users' behalf without being asked each time.</description><category>Manus</category></item><item><title>OpenClaw finally ships a real Linux desktop app</title><link>https://theagentlog.com/updates/2026-09-openclaw-linux-desktop/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-openclaw-linux-desktop/</guid><description>OpenClaw 2026.8.2 came out on September 1, 2026. It adds Quick Chat, a small chat bar you can summon with Ctrl+Shift+Space on X11 or from the tray icon (global shortcuts aren't available on Wayland, so it's tray-only there), a real Linux app instead of just a browser tab, and safer updates, so your settings won't get wiped and a bad update won't break a running agent.</description><category>OpenClaw</category></item><item><title>Perplexity Computer splits work between the cloud and your Mac</title><link>https://theagentlog.com/updates/2026-09-perplexity-computer-hybrid-compute/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-09-perplexity-computer-hybrid-compute/</guid><description>On September 1, 2026, Perplexity added hybrid compute to its Mac app. A single task can now split between cloud models, which handle reasoning and web search, and a local model on your Mac, which handles private files and sensitive information using an on-device privacy filter Perplexity open-sourced. It needs Apple silicon, macOS 15 or later, and at least 24GB of memory, and is only available on Pro, Max, and Enterprise plans.</description><category>Perplexity Computer</category></item><item><title>Grok Bot adds plugins for Outlook, Calendar, and OneDrive</title><link>https://theagentlog.com/updates/2026-08-grok-bot-microsoft-plugins/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-grok-bot-microsoft-plugins/</guid><description>On August 31, 2026, xAI gave Grok Bot new plugins that let it read, write, and act across Outlook, Calendar, and OneDrive. This extends the same always-on agent capabilities Grok Bot already had elsewhere to Microsoft 365 accounts specifically.</description><category>Grok Bot</category></item><item><title>Hermes Agent v0.21.0 adds Bot Mode, agent-to-agent group chats</title><link>https://theagentlog.com/updates/2026-08-hermes-agent-v0-21-0/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-hermes-agent-v0-21-0/</guid><description>Nous Research released Hermes Agent v0.21.0 on August 31, 2026. The big new feature is &quot;Bot Mode&quot;: named agents that stay running and can chat with each other in groups. It also adds scheduled tasks that remember past context, live control over sub-agents, and a dashboard for managing all of it. This follows several smaller updates through August.</description><category>Hermes Agent</category></item><item><title>Miso launches its AI travel agent in iMessage</title><link>https://theagentlog.com/updates/2026-08-miso-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-miso-launch/</guid><description>Miso's founder Martin Mrozowski announced on August 31, 2026 that the company's AI-native travel agent is live at try.miso.com, after about a year and a half of building it. You text it inside iMessage and it remembers your preferences (seat choice, points balances, hotels you've liked before) so you don't have to explain them again on every trip. It handles flights and hotels, and the company says some users have already booked more than $100,000 in trips through it.</description><category>Miso</category></item><item><title>OpenClaw 2.0 ships masked credentials and session permission controls</title><link>https://theagentlog.com/updates/2026-08-openclaw-2-0/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-openclaw-2-0/</guid><description>OpenClaw shipped version 2026.8.1, called OpenClaw 2.0, on August 31, 2026. It's the biggest release yet, with 933 contributors and over 16,000 merged changes after a two-month pause in updates. Among the highlights: your agent can now request a credential through a masked prompt that never puts its value in chat or model context, new session permission modes control filesystem access, sessions can now run on paired devices or cloud workers so you can collaborate, you can search past conversations by exact words or phrases, and the release includes general startup speed improvements.</description><category>OpenClaw</category></item><item><title>Soar launches its AI travel agent after a year in beta</title><link>https://theagentlog.com/updates/2026-08-soar-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-soar-launch/</guid><description>Soar came out of beta and launched in 2026 after about a year of testing, according to PhocusWire's Startup Stage series. It's an AI travel agent you text: it connects to your Gmail to automatically track every flight, hotel, and Airbnb booking and build a traveler profile from your inbox, then helps book new trips, checks you in for flights, handles changes, and gives in-trip support over iMessage. The company is backed by former Expedia CEO Erik Blachford.</description><category>Soar</category></item><item><title>Grok Bot can now read and post on X</title><link>https://theagentlog.com/updates/2026-08-grok-bot-x-integration/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-grok-bot-x-integration/</guid><description>On August 29, 2026, xAI connected Grok Bot directly to X. Linking your X account gives the bot the ability to search posts, read timelines, check mentions, and summarize activity. Paid Grok Bot users also get free starter X API credits. xAI calls this a first version with more capability planned.</description><category>Grok Bot</category></item><item><title>Grok Bot now lets you copy and share other people's bots</title><link>https://theagentlog.com/updates/2026-08-grok-bot-templates/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-grok-bot-templates/</guid><description>On August 28, 2026, xAI added template sharing to Grok Bot. You can turn a bot you have set up over time into a link. Anyone who clicks the link and adds it gets their own copy of its instructions, skills, routines, and connected integrations. Your own logins, computer access, and chat history are not shared.</description><category>Grok Bot</category></item><item><title>Cowork gets its own browser, no more Chrome extension needed</title><link>https://theagentlog.com/updates/2026-08-claude-cowork-built-in-browser/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-claude-cowork-built-in-browser/</guid><description>On August 26, 2026, Anthropic built a browser directly into Claude Cowork. When a task needs the web, Cowork now opens a custom browser in a side panel instead of driving Chrome through the Claude in Chrome extension. It's live on Enterprise plans and rolling out to Pro, Max, and Team subscribers, on by default, with an option to switch back to the extension in settings. The browser has no access to your logins for email, banking, or other services unless you enter them manually, so it's built for fetching public information rather than acting as you. Anthropic says prompt injection, where hidden instructions on a page hijack an agent, is still a risk and relies on the same guardrails as the Chrome extension.</description><category>Claude Cowork</category></item><item><title>Grok Bot is now included with every SuperGrok and Cursor plan</title><link>https://theagentlog.com/updates/2026-08-grok-bot-all-plans-expansion/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-grok-bot-all-plans-expansion/</guid><description>On August 26, 2026, xAI expanded Grok Bot to every SuperGrok and Cursor tier: SuperGrok, SuperGrok Plus, SuperGrok Heavy, Cursor Pro, Cursor Pro+, Cursor Ultra, and both Cursor Teams plans (Standard and Premium). Grok Bot comes with its own separate usage allowance, so using it doesn't eat into your regular Grok or Cursor usage. This builds on the free trial xAI added on August 21.</description><category>Grok Bot</category></item><item><title>Instinct raises $250M Series B at a $2.5B valuation</title><link>https://theagentlog.com/updates/2026-08-instinct-series-b/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-instinct-series-b/</guid><description>Instinct raised $250 million in a new funding round led by Index Ventures and Benchmark. That brings its total funding to $350 million and values the company at $2.5 billion, about five times higher than just a few weeks ago. It's still invite-only.</description><category>Instinct</category></item><item><title>Claude's memory now works the same in chat and in Cowork</title><link>https://theagentlog.com/updates/2026-08-claude-cowork-shared-memory/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-claude-cowork-shared-memory/</guid><description>On August 25, 2026, Anthropic merged the memory system behind Claude chat and Claude Cowork. What Claude learns about you in one now carries over to the other, so handing off a task to Cowork no longer means re-explaining context you already gave it in chat. You can see everything Claude remembers, topic by topic, and edit or delete any of it. Sensitive topics like health, race, or political beliefs stay off by default unless you turn them on. It's on by default on Free, Pro, and Max plans, across web, desktop, and mobile.</description><category>Claude Cowork</category></item><item><title>Meta's Hatch could be the next OpenClaw rival, if it ships</title><link>https://theagentlog.com/updates/2026-08-hatch-leaked-plans/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-hatch-leaked-plans/</guid><description>Leaked Meta documents describe a new AI agent called Hatch. It would understand what you want and complete multi-step tasks for you inside Instagram and WhatsApp, starting with shopping, then later working with Outlook, Yelp, Reddit, Etsy, and DoorDash. Reports say the top plan could cost up to $199.99 a month. Meta hasn't confirmed a launch date.</description><category>Muse</category></item><item><title>Perplexity's agent can now run free on your own Nvidia GPU</title><link>https://theagentlog.com/updates/2026-08-perplexity-computer-portable/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-perplexity-computer-portable/</guid><description>On August 25, 2026, Perplexity launched Portable Computer, a version of its Computer agent that runs entirely on hardware you already own instead of the cloud. It works on Nvidia's DGX Spark desktop and on Linux machines with Nvidia RTX GPUs, built in partnership with Nvidia. Because it runs locally, tasks don't cost any tokens.</description><category>Perplexity Computer</category></item><item><title>Wajo launches Fo, an action agent with its own phone, email, and card</title><link>https://theagentlog.com/updates/2026-08-wajo-fo-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-wajo-fo-launch/</guid><description>Wajo launched Fo on Product Hunt around August 25, 2026, calling it its first &quot;action agent&quot; for everyone. Fo gets its own identity: its own email address, phone number, and credit card, so it can join group texts, make calls, schedule appointments, change flights, place orders, and send gifts without you handing over your own card details. Wajo describes this as built to &quot;empower, not impersonate.&quot; There's a free tier that includes chat, email, text, and voice access with an agent card, and a $99 a month Pro plan that adds the ability to build your own Fo-like agents, capped at 100 tasks a day. Wajo is led by Shivani Poddar, who previously worked at Google DeepMind and on the Jules agent, and the company is based in San Francisco. Access has been through a waitlist.</description><category>Fo</category></item><item><title>Instinct kept reading your Gmail hours after you disconnected it</title><link>https://theagentlog.com/updates/2026-08-instinct-privacy-scrutiny/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-instinct-privacy-scrutiny/</guid><description>In late August, testers found that Instinct kept reading and storing Gmail messages even three hours after they had disconnected it, and it saved the messages as plain, unprotected text. Its terms of service also let Instinct keep a permanent license to use your data. A researcher even showed that fake instructions hidden inside an email could trick Instinct into doing things it shouldn't. Instinct added a manual &quot;delete my data&quot; button after this came out, and said it takes the concerns seriously.</description><category>Instinct</category></item><item><title>Grok Bot opens to cheaper plans, adds a free trial</title><link>https://theagentlog.com/updates/2026-08-grok-bot-beta-expansion/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-grok-bot-beta-expansion/</guid><description>On August 21, 2026, xAI opened up Grok Bot access beyond the original expensive tiers (SuperGrok Heavy, Cursor Ultra and Teams Premium) to also include SuperGrok Plus and Cursor Pro+ subscribers. It also added a free trial with limited usage for everyone, on Mac, iOS, Windows, and Linux.</description><category>Grok Bot</category></item><item><title>xAI ships Grok Bot, a $120-$300/mo always-on AI coworker</title><link>https://theagentlog.com/updates/2026-08-grok-bot-beta-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-grok-bot-beta-launch/</guid><description>xAI launched Grok Bot in beta on August 11, 2026. Each Grok Bot user gets one persistent cloud computer, shared by every Bot that user runs, complete with its own browser, files, and terminal, so it keeps working even after you close your laptop. It signs into your other tools using its own login. You need a paid plan to use it: SuperGrok Heavy ($300/month), Cursor Ultra ($200/month), or Cursor Teams Premium ($120/seat/month). Since launch, xAI has kept adding features like shared templates and easier account switching.</description><category>Grok Bot</category></item><item><title>Manus splits from Meta, force-deleted some users' data for 48 hours</title><link>https://theagentlog.com/updates/2026-08-manus-meta-separation/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-manus-meta-separation/</guid><description>On August 11, 2026, Manus said it was splitting from Meta, which had bought it back in December 2025. To follow certain regulations, Manus deleted some users' data from the Meta period and paused access from August 23-25, though it gave people tools to back up their data first. Manus says new features are on the way.</description><category>Manus</category></item><item><title>Lindy kills its AI agent and relaunches as Lindy Teammate, a Slack coworker</title><link>https://theagentlog.com/updates/2026-08-lindy-teammate-pivot/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-lindy-teammate-pivot/</guid><description>On August 10, 2026, Lindy founder and CEO Flo Crivello announced &quot;we kill the AI agent&quot; and introduced Lindy Teammate: an AI coworker that works out of Slack instead of iMessage or SMS. You add it to your team's Slack channels, and anyone can hand it work directly there. It sits in on meetings, connects to the tools your company uses, and keeps a running memory of context from the messages and meetings around it. This is Lindy's third distinct product in about 18 months: a no-code agent builder starting in 2024, then Lindy Assistant, a personal iMessage and SMS assistant, from February 2026, and now Lindy Teammate.</description><category>Lindy Teammate</category></item><item><title>Pally raises $5.2M seed for an AI assistant that lives in your texts</title><link>https://theagentlog.com/updates/2026-08-pally-seed-funding/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-pally-seed-funding/</guid><description>Pally raised $5.2 million in seed funding at a $30 million valuation, led by Cyber Fund and Y Combinator, with Pioneer Fund, Founders Inc, 468 Capital, and angel investors also joining. Pally is a text-based AI assistant: you message it on iMessage or Android RCS, and it connects to Gmail, Outlook, Slack, and Notion to manage your inbox, book flights, reply to messages, and make reservations. It launched in June 2026 and was founded by Haz Hubble and Wyatt Lansford.</description><category>Pally</category></item><item><title>Genspark open-sources GenOffice, bakes Super Agent into docs</title><link>https://theagentlog.com/updates/2026-08-genspark-super-agent-genoffice/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-08-genspark-super-agent-genoffice/</guid><description>Genspark released GenOffice Alpha, a free office suite (Docs, Sheets, Slides, PDF) for Mac and Windows. It's open source, and Genspark Super Agent is built right into the editor instead of living in a separate chat window. Basic editing is free; AI actions like research and slide-making use paid credits.</description><category>Genspark Super Agent</category></item><item><title>Gemini Spark can now browse Chrome on your desktop for you</title><link>https://theagentlog.com/updates/2026-07-gemini-spark-chrome-auto-browse/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-07-gemini-spark-chrome-auto-browse/</guid><description>On July 30, 2026, Google gave Gemini Spark the ability to use desktop Chrome directly, browsing with your logged-in accounts and saved passwords instead of only a separate remote browser. Google's example tasks include scheduling apartment viewings you've saved or researching and starting a flight booking. You still have to approve payments and other sensitive actions, and the browser has protection against prompt injection. The Chrome integration is rolling out first in the US. The same day, Google also brought Gemini Spark to Google AI Pro subscribers in more than 160 additional countries, following a full US rollout.</description><category>Gemini Spark</category></item><item><title>Cognition buys Poke, the text-message AI agent, for $100M+</title><link>https://theagentlog.com/updates/2026-07-poke-cognition-acquisition/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-07-poke-cognition-acquisition/</guid><description>On July 24, 2026, Cognition, the company behind the AI coding agent Devin, bought Poke's maker, The Interaction Company of California, in a deal worth low nine figures. Poke lets people text an AI agent like a friend over iMessage, SMS, Telegram, and WhatsApp, and became the first agent approved on Apple's Messages for Business platform back in June. It had handled over 100 million messages in the three months before the deal.</description><category>Poke</category></item><item><title>Claude Cowork expands beyond desktop to web and mobile</title><link>https://theagentlog.com/updates/2026-07-claude-cowork-web-mobile/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-07-claude-cowork-web-mobile/</guid><description>On July 7, 2026, Anthropic expanded Claude Cowork beyond desktop to web and mobile, in beta. Until then, Cowork only lived on your laptop, so a task stopped when you stepped away. Now your sessions and files follow you: start a task at your desk, check on it from your phone, and pick up the finished output anywhere. Scheduled tasks run even with no device online. When Claude reaches a decision only you can make, the question reaches your phone, and nothing ships until you've reviewed and approved it. Desktop remains the full experience, since it can reach folders and apps on your own computer that web and mobile can't.</description><category>Claude Cowork</category></item><item><title>ChatGPT Workspace Agents starts charging for usage</title><link>https://theagentlog.com/updates/2026-07-chatgpt-workspace-agents-billing/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-07-chatgpt-workspace-agents-billing/</guid><description>The free preview for ChatGPT Workspace Agents ended on July 6, 2026 (OpenAI had already pushed the original May 6 cutoff back once). Every agent run invoked inside ChatGPT Business, Enterprise, Edu, or Teachers workspaces now draws down credits on top of the per-seat subscription these teams already pay, based on input, cached input, and output tokens. Runs triggered outside ChatGPT, such as an agent replying in a connected Slack channel, aren't metered yet.</description><category>ChatGPT Workspace Agents</category></item><item><title>Town launches out of beta with a $55M Series A</title><link>https://theagentlog.com/updates/2026-06-town-series-a-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-06-town-series-a-launch/</guid><description>Town came out of beta on June 3, 2026, with a $55 million Series A led by Andreessen Horowitz (a16z), joined by Forerunner Ventures, First Round, Alt Capital, and Conviction. Town is a personal AI assistant, called a Townie, that connects to your email, calendar, Slack, docs, and WhatsApp, learns how you work, and starts proactively helping with tasks like scheduling, follow-ups, and admin work. It was founded by Jean-Denis Grèze, former CTO of Plaid, and Tony Vincent, former director of applied AI at Google.</description><category>Town</category></item><item><title>Windsurf is now Devin Desktop</title><link>https://theagentlog.com/updates/2026-06-devin-desktop-rebrand/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-06-devin-desktop-rebrand/</guid><description>Cognition shipped Windsurf as Devin Desktop on June 2, 2026, as a standard over-the-air update. Plans, pricing, settings, extensions, keybindings, and MCP connections all carried over automatically, no action needed. The default launch surface changed from a code-first editor to an Agent Command Center, a Kanban board of every local and cloud agent you're running, sorted by status. Devin Desktop also added support for the open Agent Client Protocol, so Codex, Claude Agent, Gemini CLI, and other agents can run inside the editor alongside Devin. Cascade, the local agent that powered Windsurf, is being replaced by Devin Local, a Rust rewrite Cognition says is more token-efficient and supports parallel subagents. Cascade stays available only until July 1, 2026.</description><category>Devin Desktop</category></item><item><title>Sonder co-founder launches Odessia, an AI travel agent that books your trip in one chat</title><link>https://theagentlog.com/updates/2026-06-odessia-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-06-odessia-launch/</guid><description>Francis Davidson, co-founder of the now-collapsed Sonder, launched Odessia on June 2, 2026. You describe a trip and it plans a personalized itinerary, then books the flights, hotels, and activities in the same chat instead of sending you to a site like Expedia to finish. The company raised $6 million in seed funding led by Sequoia Capital. Odessia is currently in preview, with the company asking for user feedback.</description><category>Odessia</category></item><item><title>Google launches Gemini Spark, a 24/7 personal AI agent</title><link>https://theagentlog.com/updates/2026-05-gemini-spark-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-05-gemini-spark-launch/</guid><description>At Google I/O 2026 on May 19, 2026, Google introduced Gemini Spark, a 24/7 personal AI agent built into the Gemini app. Give it a task and it keeps working in the background around the clock, even when your phone and laptop are off, running on dedicated virtual machines on Google Cloud and powered by Gemini 3.5 with Google's Antigravity agent harness. It connects to Gmail, Calendar, Drive, Docs, and Sheets, and checks with you before taking major actions. It launched to Google AI Ultra subscribers the week after the announcement, alongside a macOS app integration.</description><category>Gemini Spark</category></item><item><title>Mindtrip launches agentic AI flight booking with Sabre and PayPal</title><link>https://theagentlog.com/updates/2026-05-mindtrip-agentic-flights/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-05-mindtrip-agentic-flights/</guid><description>On May 6, 2026, Mindtrip launched an agentic AI flight booking feature built with Sabre and PayPal. It lets you search, select, and pay for a flight entirely inside Mindtrip's chat interface, using Sabre's Mosaic travel platform for real-time flight data and PayPal's agentic commerce tools for payment. Mindtrip later expanded the same approach to hotels with Mindtrip Stays.</description><category>MindTrip</category></item><item><title>Caddy launches a personal AI that lives in your texts</title><link>https://theagentlog.com/updates/2026-04-caddy-public-beta/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-04-caddy-public-beta/</guid><description>Caddy is a proactive personal AI you message directly, with no separate app to open: on iPhone it works over iMessage, on Android over RCS and SMS. It watches your calendar, messages, and tasks, and texts you when something needs attention. It can also handle reminders, follow-ups, and calendar updates in the background. The company launched to public beta on April 30, 2026, and it's free while in beta. Caddy is part of Y Combinator's Fall 2025 batch, based in New York, and was founded by Connor Waslo and Rajiv Sancheti, who previously worked together on Loom's AI suite. Early reports put its seed funding at around $500,000, with Pioneer Fund among the investors.</description><category>Caddy</category></item><item><title>OpenAI launches Workspace Agents, shared AI teammates for teams</title><link>https://theagentlog.com/updates/2026-04-chatgpt-workspace-agents-launch/</link><guid isPermaLink="true">https://theagentlog.com/updates/2026-04-chatgpt-workspace-agents-launch/</guid><description>OpenAI launched Workspace Agents on April 22, 2026, an evolution of custom GPTs built for teams instead of individuals. You describe a recurring workflow or upload a file, and ChatGPT turns it into a shared agent with its own workspace, files, tools, and memory. Agents run on Codex in the cloud, keep working after you close ChatGPT, pull context from connected apps like Slack, Google Drive, and Salesforce, and ask for approval before sensitive actions like sending an email. It launched as a research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans, free until May 6, 2026, after which credit-based billing applies.</description><category>ChatGPT Workspace Agents</category></item></channel></rss>