OpenClaw 2.0 Launches With Usability Boost But Security Concerns Persist

The Year of the Claw: OpenClaw Rolls Out New Version of Its Viral AI Agent - CNET

Earlier this year, I went on record saying that 2026 would be the year of the AI agent. With the seemingly never-ending rollout of agentic AI since then, I feel validated in that prediction. However, I may make one amendment to that statement now: It's the year of the claw. I first realized agents would be the next big thing in AI when OpenClaw, an open-source AI agent that runs right on your devices and can do a slew of tasks independently on your behalf, went absolutely viral in January. The creator, Peter Steinberger, even joined OpenAI in February to "work on bringing agents to everyone." Now, OpenClaw is regenerating into a newer, more capable lobster. OpenClaw recently announced the rollout of OpenClaw 2.0, the "largest update" in its history, bringing a number of fixes, new features and capabilities. Steinberger once said, "Molting is what lobsters do to grow." And it seems like OpenClaw is shedding its shell and molting to be bigger and better than before. What is OpenClaw? OpenClaw is an AI assistant that doesn't just chat, but executes tasks on your behalf without any oversight. It can be integrated with the apps you use most often, like WhatsApp, Telegram, iMessage, Slack, Discord, Signal and more, to operate like a personal assistant that never stops to eat, sleep or be human. It also doesn't need specific hardware to run, making it accessible on really any device you have. The assistant went viral across social media, Reddit and internet spaces that are all about AI. Part of the allure was that OpenClaw is open-source, meaning it's freely accessible for you to use, personalize and share, but also because of its chaotic origin story. I already wrote in-depth about the hectic beginnings of OpenClaw, which included multiple renames from Clawdbot to Moltbot to OpenClaw, and broke the internet... twice. The real catch is that while OpenClaw was a huge hit for being capable, autonomous and powerful, it came with equally concerning security risks that were scaling just as fast. (You can read my full analysis of OpenClaw to learn about these security concerns.) The biggest worry about AI agents is their proclivity to go rogue. These agents will often work to accomplish tasks by any means necessary, even going beyond the settings and boundaries set by the humans who created them. This means they can access the entire internet with minimal effort, prompting experts to call AI agents "cybersecurity nightmares." Despite warnings from cybersecurity experts, OpenClaw has been a global success, even sparking numerous copycats. Many AI companies have spent the better part of a year trying to replicate OpenClaw's success with the rollout of Gemini Spark, Claude Cowork, ChatGPT Work, Grok Bot and Copilot Studio. What is new in OpenClaw 2.0? OpenClaw 2.0 emerged from the goal of "simpler setup and a first-class browser experience," according to a blog post about the launch by the OpenClaw Foundation's Hannes Rudolph. The 2.0 version includes updates to "every part of OpenClaw" from installation, messaging, memory, skills, the browser, plugins, security and numerous fixes, which combined to create a new, revamped experience. The first major update to the installation process should make it easier for first-time OpenClaw users to get started. The changes include starting the setup with "what is already on someone's computer," like your existing ChatGPT or Claude subscriptions, API keys and local models. From there, configuration was simplified to allow you to get prompting quicker and finalze set up later. The other big change comes to the browser app, which was rebuilt to open directly into a conversation with your claw (the name for your personal AI assistant in OpenClaw). Many people use OpenClaw because it's genuinely capable. The AI assistant can easily fit into your most-used tools, such as calendar, notes and email, so it feels more like a part of your daily routine and less like autonomous software. If you want to try OpenClaw for yourself, you can visit openclaw.ai for documentation, installation guides and security checklists. If you do decide to take the chance on OpenClaw, make sure you're aware of the security risks and the potential for agents to go rogue - and maybe consider using a spare laptop.

OpenClaw 2.0 pours glitter on slow-burning security dumpster fire

OpenClaw has unveiled what its makers call its largest ever update - large enough to earn a 2.0 moniker - with usability taking center stage, along with some security updates that critics are suggesting will be insufficient. The OpenClaw foundation announced the release of version 2.0 of its AI agent harness on Sunday, describing it as something with far more scope than they ever intended it to have. "This update touches every part of OpenClaw," Foundation community manager Hannes Rudolph said of the update. "We started by simplifying installation and rebuilding the browser app as a first-class experience, but doing that properly meant carrying the cleanup through the rest of OpenClaw until it became OpenClaw 2.0." It's those two features - the rebuilt installation experience and redesigned interface - that Rudolph dedicates most of his announcement to. OpenClaw is an open-source, self-hosted AI agent harness that allows users to build their own AI agents and connect them to whatever apps and services they want. OpenClaw went viral shortly after its launch due to its extensive capabilities, and helped launch the AI agent craze. But, by empowering AI models with agentic capabilities, it exposed numerous security problems with unrestrained automation. In version 2.0, the new installation process is designed to be simpler, ostensibly to get more people using OpenClaw. "We cut or simplified a lot of configuration and moved the rest out of initial setup, letting people get to a first conversation faster and finish setting up their Claw by talking to it," Rudolph explained. As for the user experience, Rudolph explained that the OpenClaw browser app has been redesigned into "a first-class experience" where users can continue setup and interact with their agent. "The web-based experience in OpenClaw now feels more familiar to anyone who uses apps like ChatGPT, Claude, Gemini, or Perplexity, with conversations in the sidebar and the one you are working in at the centre instead of opening on a separate Overview page," the patch notes for the release explains. In other words, OpenClaw's basic interface now looks just like the chat interface for every other AI service you've likely used on the web. The last major feature update added in OpenClaw 2.0 is shared cloud sessions. Per the announcement, OpenClaw previously had no way to include multiple team members in a single instance without the Claw involved losing its memory. Shared cloud sessions correct that, enabling multiple people to interact with a single Claw while context is maintained across users and a continuous chat, giving OpenClaw feature parity with the agent harnesses offered by frontier labs like Anthropic and OpenAI, which allow collaboration for enterprise users. What about security? Since launching in November 2025, OpenClaw has deservedly earned its reputation as a complete security mess - not only in the code itself but for users and those who are unfortunate enough to come in contact with a Claw's orders as well. Celebrity UK mathematician Professor Hannah Fry tested OpenClaw out earlier this year, finding it was ready and willing to share her private information when threatened. In another instance, an OpenClaw agent hacked a gym's waiting list and forced its user into a full class, displacing other reservations, when simply asked to get him on the list. So, what is OpenClaw doing to improve on these risks as part of the update? Not that much, based on a reading of the patch notes. Shared sessions, for example, are a great way to introduce collaborative Claws at work, but the OpenClaw foundation states in the patch notes that the shared session controls "are not tenant isolation or a security boundary." In other words, you'd better be sure there's no need to isolate various OpenClaw instances. A new protected credentials feature has been added that allows users to share credentials with agents in shared environments without exposing them in chat. That's great, and as explained in the patch notes it's further secured in a local secret store that "separates Protected values from Agent-readable environment values." What's not great, on the other hand, is the fact that "Secret Store values are not encrypted at rest and depend on the filesystem permissions of OpenClaw's state directory." A new sandbox for contributor-controlled code was also announced, with the patch notes referring to an environment for untrusted code isolation. Again, great - except sandboxing is turned off by default. In other words, this release is doing a lot to make installing and getting OpenClaw up and running for more people, but it's not bringing security by default along with that accessibility. As we've warned before, granting a capable and potentially dangerous tool like this widespread access to your systems and credentials ought not be done lightly, fancy new wrapper or not. ®

OpenClaw

OpenClaw is an open-source AI assistant that runs on your machine and works from the chat apps you already use. OpenClaw runs on your own devices and can connect to services such as WhatsApp, Telegram, Slack, and Discord. It supports multiple AI models and can use tools and plugins to handle tasks such as managing email and calendars, browsing the web, and automating everyday workflows. Features Runs on Your Machine Mac, Windows, or Linux. Bring hosted, subscription-backed, gateway, or local models. State lives on your machine, not a vendor cloud. Chat Where You Are Talk to it on WhatsApp, Telegram, Discord, Slack, Signal, iMessage, or any of its 29 channels. Works in DMs and group chats. Persistent Memory Remembers your preferences, projects, and people from one conversation to the next. Full System Access Browses the web, fills forms, reads and writes files, runs shell commands. Full access or sandboxed-your choice. Works With Your Team Share one gateway: sessions the whole team can open and steer, live presence, and commits credited to the people behind them. Skills & Plugins Extend with community skills or build your own. It can even write its own. Integrations Connect OpenClaw to the chats, models, tools, and devices you already use. Bring API keys, sign in with supported subscriptions, route through a gateway, or run models locally. Extend the runtime with official plugins and community skills. What's New OpenClaw 2.0, Accidentally (AKA v2026.8.1) Today we released by far the largest update in the history of OpenClaw. It was built by 933 contributors, including 569 first-time contributors, and is composed of over 16,000 pull requests. You can read the full release notes HERE. This update touches every part of OpenClaw, including installation, messaging, memory, skills, models, automations, the browser and native apps, plugins, security, and a very long tail of fixes. We started by simplifying installation and rebuilding the browser app as a first-class experience, but doing that properly meant carrying the cleanup through the rest of OpenClaw until it became OpenClaw 2.0. Why this took nearly two months Before this update, we had shipped 106 releases in 230 days, most within a day or two of the one before them, so going nearly seven weeks without shipping was not normal for us. The release cadence slowed, but development moved in the opposite direction because our team was growing, and the increased volume and pace of work outgrew both the foundation of OpenClaw and the process we used to ship it, so we reworked both at the same time. That acceleration left us with a release containing roughly 50% of all pull requests ever merged into OpenClaw, so we took the extra time to make sure it worked for people starting from scratch and people upgrading an existing Claw rather than shipping quickly and handing them an update that broke what they already had. Getting to a useful Claw faster Installation was the first place we needed to make OpenClaw easier, so for first-time installs we are starting with what is already on someone's computer, including existing ChatGPT or Claude subscriptions, API keys, and local models. We cut or simplified a lot of configuration and moved the rest out of initial setup, letting people get to a first conversation faster and finish setting up their Claw by talking to it. The browser app is where most people meet OpenClaw and have their first conversation, so we rebuilt it as a first-class experience where they can keep setting things up, return to ongoing work, or follow along live when they want to. The rebuilt browser app opens directly into a conversation with your Claw. How a useful Claw can grow Your Claw does not need to be complicated, and a simple workflow might have it watch your inbox for your kids' school emails and send you a Telegram message whenever something important comes through, like homework due or an upcoming activity you need to prepare for. It uses one inbox, looks for a few important things, and sends the result to one place, which is already enough to be useful. From there, a task can reach across more places without becoming harder to use, so when your brother sends an iMessage asking which iPad you bought for your dad, you can skip searching your email for the receipt and simply tell your Claw that your brother just messaged, then ask it to find the answer and send it to him. The Claw just does it. The same progression showed up inside our own team while building this release as we used our Claws for more of the work and wanted to share tasks, collaborate on them, and sometimes hand them off entirely. OpenClaw had no way to bring another team member into the work without losing what the Claw already knew. Shared cloud sessions changed that and turned OpenClaw into a multiplayer experience our team now uses to build OpenClaw, bringing the right person into live work or handing it over with the context intact. A user-built dashboard inside a shared multiplayer OpenClaw workspace. What this adds up to Your OpenClaw starts with one useful workflow and grows as far as you want it to, with the possibility of reaching across more of your life and work or becoming multiplayer with your family or team when you want and how you want. This has changed the relationship people have with software, turning it from something designed somewhere else that you have to accept into something you can tell what to do, shape around your life and work, and actually own.

OpenClaw is yesterday's hype and today's AI blueprint

Roughly eight months after its astonishing debut, the OpenClaw hype has largely faded, and not even the release of OpenClaw 2.0, complete with bolstered security and easier user onboarding, is likely to vault OpenClaw back into the limelight. But while OpenClaw isn't on everyone's lips anymore, the arrival of its new and improved version serves as a reminder of how much it changed the game, ushering in an era of agentic AI that touches everything from ChatGPT Work and Claude Cowork to Google's Gemini Spark and Nvidia's own "claw"-inspired products. As for OpenClaw 2.0 itself, the new release is billed as being much easier for first-time users, allowing it to quickly check which AI models and services you have on hand, including ChatGPT or Claude subscriptions, API models, and local LLMs. The OpenAI-backed OpenClaw (which continues to be both open-source and free) now skips over "a lot" of the most confusion customization steps in favor of helping users to "get to a first conversation faster," as well as to "finish setting up their Claw by talking to it." The OpenClaw browser is now a "first-class experience," according to OpenClaw's press release, while several of the security holes that -- quite rightly -- scared off new users have been fixed. Security credentials are now masked so they never appear in chat, for example, while enhanced workspace restrictions make it harder for agents to wander all over your system. Perhaps most importantly, third-party plugins on the ClawHub must now show you what they do and where they come from before you install them. But while the new OpenClaw promises to be both easier and more secure, it's still a long way from being as safe (well, relatively safe) as ChatGPT. Despite its new security features, OpenClaw still gets practically unfettered access to your PC or Mac, allowing it to fire off shell commands, access your files, and snap screenshots. A warning during the onboarding process makes it plain: "Only enable OpenClaw if you understand the risks and trust the prompts and integrations you use." Personally, I still wouldn't recommend OpenClaw for everyday AI users, and I remain hesitant to install it myself.

OpenClaw 2.0: How to try the updated AI agent, what's new

A major new version for the open source AI assistant platform OpenClaw is here. Credit: OpenClaw A new version of OpenClaw, the open-source AI agent, is available now. OpenClaw 2.0 was released over the weekend, promising a "simpler setup" and a new browser app. On its website, OpenClaw calls it "by far the largest update in the history of OpenClaw." The OpenClaw Foundation says that OpenClaw 2.0 was completely crowdsourced with 933 contributors. Of those 933 contributors, 569 of them contributed to the project for the first time. * Guess what Apple will release in September in our Big Guessing Game: Apple edition. Enter now for a chance to win an Apple Watch! Here's what we know about OpenClaw 2.0, including how you can try it for yourself. What's new with OpenClaw 2.0 The new release acknowledges that the original OpenClaw (the AI agent formerly known as Clawdbot, then Moltbot) was more complicated than it needed to be for a first-time users. The update notes say that configurations were moved out of the installation process, simplified, or completely removed in order to make the installation process quick and easy. The update also includes a completely rebuilt browser in order to simplify OpenClaw. The browser is where users interact with OpenClaw, assign tasks to agents, and track ongoing work and processes. OpenClaw 2.0 also introduces Shared Cloud Sessions, a feature labeled as a "multiplayer experience." Shared Cloud Sessions allow a user to bring in team members onto ongoing work, allowing multiple people to come in and take over existing tasks with all of the context still attached. Other new features in OpenClaw 2.0 include a new search function for previous conversations, live session cards that display task progress in real-time, and an assortment of additional new widgets and dashboard features. There are a number of major security updates in OpenClaw 2.0 as well, including a new private credential request feature, which allows users to securely share credentials with an agent without exposing them in a chat. OpenClaw quickly went viral when it was released late last year. The self-hosted, open-source agentic AI platform developed such a cult following that founder Peter Steinberger was quickly hired by OpenAI. OpenClaw is now managed by a nonprofit that's supported by OpenAI. However, OpenClaw has struggled to maintain its position as other AI platforms rolled out similar agentic AI features that were more easily accessible to non-technical users. Perhaps a simplified setup process in this newest version will help OpenClaw retain its positioning among AI users. How to try OpenClaw 2.0 Since OpenClaw is open-source, developers should have no problem getting it up and running. For less technically-inclined users, you can download OpenClaw 2.0 for Mac or Windows at the OpenClaw website. Simply download the installer and follow the instructions to get started. You can also read all about OpenClaw v2026.8.1 (aka OpenClaw 2.0) in the official release notes. Be warned -- you should understand the cybersecurity risks of handing your device over to an AI agent before installing or using OpenClaw. Disclosure: Ziff Davis, Mashable's parent company, in April 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.

OpenClaw just gave everyone version 2.0 and it comes with major updates

OpenClaw just gave everyone version 2.0 and it comes with major updates OpenClaw, the popular open-source personal agentic artificial intelligence framework, today announced the launch of version 2.0, the largest update in the agent's history. Unlike a traditional chatbot like ChatGPT, which lives in a browser tab and only answers questions, OpenClaw acts like a digital secretary that can actually execute tasks on a user's behalf and lives on their computer or in the cloud. It is a free, open-source AI agent that many users have turned into a coding agent or personal assistant to organize their inbox, manage their calendar and check in on their everyday activity. Although the version release is 2026.8.1, the OpenClaw team is calling this 2.0 because it is a massive update including mostly quality-of-life updates, better persistence, easier installation and increased security. Up front, the OpenClaw team streamlined the installation experience so that new and veteran users can detect what's already on their personal computer or cloud instance. For example, active ChatGPT or Claude subscriptions, application programming interface credentials and locally hosted models. The team said certain configuration options are set to defaults so users can start talking to their agent sooner, instead of getting bogged down guessing what's available. The dashboard has also been updated to become a real control center. Chat is now front and center and looks immediately like other AI interfaces users expect. With a text box that leads into a conversation. It allows users to start a new chat, rejoin ongoing work and watch an AI agent doing live work. OpenClaw has also rebuilt the interface to be more interactive with new widgets and dashboards. Users can bring widgets into chats, pin them to dashboards and assign actions that can be triggered by events. In particular, the team said it removed one previous limitation. Groups couldn't work together with the same AI agent. Now, shared cloud sessions change that: team members can join ongoing sessions and continue work where it left off, assign tasks to teammates with full context, and collaborate in real time on content and automations. The OpenClaw team said it now uses this "multiplayer" collaborative feature to build the AI agent framework code itself. Other quality-of-life updates include an updated search that lets users look through past conversations using exact text, words, or phrases, and reopen the surrounding messages. This will let users easily find past work and resume where they left off. Security updates include explicit trust boundaries, least-privilege controls and stronger protection against prompt abuse. OpenClaw is now aligned around who can trigger it, what it can access, where code executes and where credentials can be sent. It does this by controlling the gateway, providing permissions based on role and tool handling. Early OpenClaw versions and installations suffered from major security flaws, so these updates help reduce the attack surface. Users are still expected to harden their installation against the open internet and maintain strong authentication if they choose to connect personal information or use tools that can execute on their cloud service or local machine. This is the first major version update to OpenClaw since OpenAI Group PBC hired Peter Steinberger, the creator of the AI agent framework.

OpenClaw 2.0 Released as Biggest Update Since Launch: What's New

OpenClaw has released the biggest update to the self-hosted open-source AI agent platform since its launch. As per the company, the update was developed by 933 contributors and includes more than 16,000 pull requests. Dubbed OpenClaw 2.0, it brings changes across installation, messaging, memory, skills, models, automations, browser and native apps, plugins and security. OpenClaw has rebuilt its browser app as a first-class experience, while the initial setup process for new users has been simplified, too. OpenClaw 2.0 Changes In a blog post, OpenClaw said that the 2.0 update affects every part of the platform. Described as the largest release in its history, it comes after nearly seven weeks of development. This is said to be a significant departure from the project's earlier release cadence. The company previously shipped 106 releases in 230 days, with most updates arriving within a two-day window. Among the most notable changes is the installation process. According to the company, new users can start with resources already available on their computers, including existing ChatGPT or Claude subscriptions, API keys, and local models. To further simplify things, some of the configurations have been moved out of the initial setup. This reportedly allows users to get to their first conversation faster and complete additional configuration later on through conversations. OpenClaw has also rebuilt the browser as a first-class experience rather than an afterthought. Through this, users can continue setting up their Claw, return to ongoing work and follow tasks in real time. Another addition is the ability to search past conversations. The company says users can search visible conversation text using exact words or phrases and reopen the surrounding messages from a matching result. Meanwhile, support for sessions beyond the Gateway allows them to continue work on paired devices or cloud workers, move a session's workspace with it and reuse machines and project seeds for subsequent cloud sessions. With new durable session progress cards, users can also follow their work in real time. This includes subagent activity and accumulating edits after reloading the web or native app. There are now widgets and dashboards that make up a more interactive interface. Users can add widgets inside chats, pin them to session dashboards, grant specific actions or network origins and export rendered views as images. For recurring tasks, users can now approve an exact operation once and allow an automation to perform it again. The update also expands media support and adds new native playback controls. As part of the improved security infrastructure, OpenClaw 2.0 will allow an agent to request credentials through a masked prompt without exposing the value in the chat or model context. Despite the extensive list of changes, OpenClaw said that the platform remains self-hosted and open-source. Developers will be able to build on it without being tied to a single AI model or provider.

OpenClaw 2.0 arrives: The update that will change the way we work forever

OpenClaw 2.0 is now available. The biggest update in the brief but striking history of the AI tool opens, according to its creators, a new chapter. OpenClaw 2 has been made possible thanks to 933 contributors, 569 of them taking part for the first time, and more than 16,000 pull requests to the project. We're looking at a huge leap in the scope and flexibility of a key tool for those of us who want AI to move from answering questions to helping us with our day-to-day tasks. Simpler installation and a browser that takes center stage in the update During installation, OpenClaw 2 detects any ChatGPT or Claude subscriptions we may have, our API keys, and the local models available. With very little extra setup, we can start chatting much sooner and even finish the rest of the setup while already chatting with Claw. An approach we've already seen in other tools, including ChatGPT's integration with Mac apps. After the initial setup, the main big new feature is the browser app. From there we can chat, continue with tasks we already have underway, and track the progress. Much like Claude's computer control feature, the most interesting thing is that here we start with a simple task or automation and build on it. We can, for example, keep an eye on our emails, retrieve information from our services, or reply to the messages that reach us through different channels. And here comes the third big new feature, because in OpenClaw 2.0 sessions let us bring in other people and share our tasks while preserving the context. It's an approach that fits with multi-agent systems like OpenAI's Astra, where multiple participants work together, in this case humans or AI agents, to complete a task. A totally new way of working, and one we'll want to explore, given its enormous potential, in great depth.

OpenClaw 2.0 is finally here: Easier setup, redesigned browser and multiplayer AI

New shared cloud sessions let multiple users collaborate with the same AI agent without losing context. OpenClaw has released its version 2.0, which the project describes as its biggest update since the launch. The open-source AI agent platform says the release was built with contributions from 933 developers, including 569 people making their first contribution and includes more than 16,000 pull requests. The update brings changes across installation, messaging, memory, skills, models, automation, browser functionality, native apps, plugins and security. OpenClaw 2.0 gets simpler setup and redesigned browser experience OpenClaw 2.0 offers a revamped installation process aimed at getting new users up and running faster. The platform can now make use of resources users may already have on their computers, including ChatGPT or Claude subscriptions, API keys and locally installed models. OpenClaw has also reduced the amount of configuration required during the initial setup, with users able to complete more of the process by interacting with their AI agent. The browser interface has also been rebuilt as a central part of the OpenClaw experience. Users can start conversations, continue existing tasks and monitor their Claw's activity from the redesigned interface. The project says the changes are intended to make it easier for users to start with a simple automated workflow and gradually expand what their agent can handle. OpenClaw adds multiplayer collaboration OpenClaw 2.0 also introduces shared cloud sessions, allowing multiple people to work with the same AI agent while retaining its existing context. The feature was developed after the OpenClaw team began using agents more extensively during the development of the release and needed a way to share ongoing tasks without starting over. The project says the new multiplayer functionality can allow users to bring colleagues or family members into an active workflow or hand tasks between people while preserving the agent's context. OpenClaw said the latest release includes roughly half of all pull requests ever merged into the project. The team delayed the release to ensure the update worked for both new installations and existing users upgrading from earlier versions.