OpenClaw

OpenClaw is the AI ​​agent that is blowing the AI ​​industry’s mind. We have tested it: Crossover 1×42

by

ChatGPT and Claude are great, but they only do things when you ask them to. OpenClaw It’s something else. It is an AI agent that takes advantage of the power of ChatGPT or Claude (or other models) He becomes your personal employee and does everything you ask of himbut in an autonomous and proactive way. This is something that the industry has been promising for years, and although some steps had already been taken in that direction with AIs that, for example, can reserve a table for you in a restaurant, OpenClaw goes a little further because you basically “give them the keys to the office”. So, when you install it on a machine (or a VPS, or a Raspberry Pi, or a Docker container, or wherever you want) you give this AI agent superpowers, because it will be able to do everything it wants on that machine. You will be able to use all the apps you have, browser included, and use all those tools to do things for you. It is, we insist, like having an employee who works for you 24 hours a day and who, if you don’t want to, will never rest. The concept is super powerful, but of course it has some buts. The most important one is security risks, and in this episode we talk about how to protect yourself so that that virtual employee doesn’t end up messing you up and causing chaos. We also have to talk about costs, because this AI agent is a true “token glutton” and you will have to be practical when choosing which models you want to use it with. We talk about all that and many more things in this episode Crossover 1×42, which serves as an introduction to a fascinating topic. Be careful, this is addictive. On YouTube | Crossover In Xataka | OpenClaw changed the rules of the AI ​​race. Technology companies already have their answer: copy it

The AI ​​race is no longer about who has the most powerful model. Who launches the easiest and safest OpenClaw

by

2026 began with an earthquake in the world of AI, and it did not come from any of the big technology companies, but from an unknown programmer and his open source project OpenClaw (formerly Clawdbot and Moltbot). Not even two months have passed and we can say that the boom of this AI agent is reconfiguring the AI ​​career, causing more and more companies to jump on the bandwagon. The last one was Perplexity. Personal Computer. a month ago, Perplexity announced Computerwhich was a cloud-based tool capable of orchestrating agents using various models. The next step is Personal Computeryour own OpenClaw. can be left running on a Mac Mini and control it from another device, such as a mobile phone, exactly the same as OpenClaw, but with a simpler interface that does not require technical knowledge. Further user-friendly. Another key aspect is that they focus on security, one of the delicate points of OpenClaw. Perplexity claims that with Personal Computer, “Every sensitive action requires your approval. Every action is logged. There’s an off switch.” At the moment Personal Computer is not available yet, but if you want to try it before anyone else you can sign up for the waiting list. NVIDIA NemoClaw. Which is the most valuable company in the world has taken good note of the success of OpenClaw and a couple of days ago they announced that they will launch their own open source platform for enterprise AI agents, they will call it NemoClaw. This announcement is also important because it places NVIDIA in a position of direct competition against companies like Anthropic, OpenAI or Perplexity. This changes its position from a hardware supplier to a software competitor. and OpenAI…The project had not even been three months old when OpenAI, not only bought it, but also hired its creator Peter Steinberger. It was not the only one who bid to achieve the viral success of the moment, Meta also tried, but OpenAI was the one that won the bid. Stenberger said the project would continue to remain “open and independent.” This case is a good example of two things: how far a person can go with a good AI idea and how difficult, if not impossible, it is to compete in an ecosystem in which the competition is some of the largest and most valuable companies in the world. David against Goliath. The agentic AI race. We spent a good part of 2025 watching AI agents take their first steps, many times with quite mediocre results. It was clear that agentic AI was getting a lot betterbut I don’t think anyone expected that the first viral hit would be carried out by an independent and open source project. OpenClaw not only succeeded, it has launched a new race in AI, one that seeks the ultimate custom AI agents. OpenClaw has two barriers to entry, on the one hand requiring certain technical knowledge and on the other security. It is a very powerful agent, but sometimes unpredictable. Hence, Perplexity is appealing precisely to improve these two aspects. We’ll see who will be next. In Xataka | Social networks were born for humans: Meta has just bought one designed for AI agents Image | Pexels

China has turned OpenClaw into a viral phenomenon. And then it has prohibited its officials from using it

by

The appearance of the AI ​​agent OpenClaw has meant that we are living in a kind of second “DeepSeek moment“They know it well in China, where its use has exploded in such a way that the Government has had to act. And it has probably done well. An absolutely viral AI. The OpenClaw project has caused a real earthquake in China. In cities like Shenzhen there are queues to physically install it and people paying for others to install it remotely or in person. The AI ​​agent is breaking all popularity records for programming projects, and for example has already surpassed two legends such as React or Linux in terms of stars awarded on GitHub, a measure of the popularity of open source projects. In just three months, OpenClaw has managed to surpass the legendary leaders of this ranking in GitHub stars: react and linux. Source: Star-History.com Solution to Chinese fragmentation. The secret of this success in the Asian giant is not based only on the curiosity of users, but also on the fact that OpenClaw provides a striking solution to an endemic problem in the country: the fragmentation of business software. With an average of 150 independent IT systems per company and 60% of them without APIs or documentation, AI integration seemed to be an insurmountable wall. OpenClaw solves the problem because you can take control of the machine, “see” buttons and text boxes, click and type in browsers, and operate as if you were a human. Tokens everywhere. That ability has turned this project into an absolute “token hole.” Unlike a conventional chatbot like ChatGPT, OpenClaw works continuously and autonomously, and it is not uncommon to see an advanced user consume 50 million tokens daily. The impact has been massive: at the end of February, Chinese models such as Kimi 2.5 or DeepSeek were already devouring 61% of the global OpenRouter tokens, a platform that allows you to easily use APIs from dozens of AI models. The fever has been such that Kimi has generated in 20 days more income than all expected by its creator, Moonshot IA, by 2025. Alarm. The problem is precisely that: when software has the ability to “see” everything that happens on a screen and execute commands by itself, the security risks are enormous. This has made the Beijing government go from enthusiasm—cities like Shenzhen offer million-dollar subsidies for their development—to a policy that is now totally restrictive. Government agencies, state-owned companies and large national banks have received urgent notices prohibiting the installation of OpenClaw in office devices and even in mobile phones that are used in this type of segments. Be careful with your data. Practically since it went viral, many have warned of the cybersecurity risks involved in using OpenClaw. An initial audit of the skills available on ClawdHub detected hundreds of them as malicious. That was the germ of the OpenClaw alliance with the Spanish cybersecurity firm VirusTotalpart of Google. The risk with this project is threefold: You have access to private data Can communicate with the outside You are exposed to untrustworthy content and attacks from prompt injection One of lime and one of sand. For large Chinese technology companies, the government’s measures are bittersweet. On the one hand, they have rushed to offer one-click OpenClaw deployments in their clouds for interested users. On the other hand, state restriction has meant that some of the AI ​​startups such as Zhipu (Knowledge Atlas Technology JSC Ltd.) or MiniMax Group have quickly fallen on the stock market for the news. China and control. There is another key element in that political movement: the loss of control. The Beijing government has already fought a battle in the past to curb the power of giants like Alibaba, and that caused the “Asian Jeff Bezos”, Jack Ma, came out very badly. An autonomous AI agent that operates outside of that government control represents a challenge to the mechanisms that China has been perfecting, especially with its Great Firewall. An uncertain future. These new restrictions pose a complex future for the project in China. The Asian giant has embraced AI more than anyone else, but the security risks in this case are so clear that limits had to be set before things got out of control. Even so, the project is Open Source, which will make it difficult for its deployment to be halted by end users and enthusiasts, no matter how much the Chinese Government wants it. Image | OpenClaw | Paul Kagame In Xataka | Every time Facebook had a competitor, it bought it: it is exactly the same thing that OpenAI is doing

If the question is how much OpenClaw is taking over, the answer is… in China they are lining up to install it

by

One of the AI ​​phenomena of 2026 is OpenClawone of the most powerful tools (and at the same time dangerous) never released. In China, just a few days ago, nearly a thousand people queued in front of the headquarters of the Chinese technology company Tencent Holdings in Shenzhen. Reason: they wanted to install OpenClaw on their computers. OpenClaw fever. In less than 100 days, OpenClaw has conquered Asia. Of the more than 140,000 AI agents based on its technology, half come from China. Tencent issued an open invitation to its headquarters in Shenzen so that company engineers could install OpenClaw completely free of charge on the computers of those who came. Nearly 1,000 people physically turned out to install the software formerly known as Clawdbot and Moltbot. Because. The enthusiasm for OpenClaw comes from the combination of several key points. China’s early adoption of new technology Push by big technology companies to implement it (Tencent, Alibaba, Bytedance) Real agentic functions, far above what any chatbot offers. It’s not just OpenClaw. ArkClarw, MaxClaw, AutoClaw… The OpenClaw fever has triggered variants of the original, with derivative projects that seek to further specialize their work. It is a perfect storm that fits with the Chinese need to meet the massive demand for AI use while consuming few tokens. This matches perfectly with what China is looking for: cheap, Open Source models prepared for large volumes of use. Yes, but. Although Chinese consumers are demanding use of OpenClaw, the country’s authorities are moving in the opposite direction. As Bloomberg reports, state-owned companies and government agencies have received notices in recent days warning them not to install OpenClaw for security reasons. In the most extreme cases, and according to anonymous Bloomberg sources, the installation of this software has been banned in critical sectors such as banking and some government agencies. Lobster farming. There is a new buzzword creating a trend in China: “locust farming”. One that has to do with the OpenClaw logo and the tendency to install its variants. The project has already surpassed Linux on GitHub in terms of ratings, and despite doubts about its security, it has made something clear: agentic AI is not so far away from what we thought. In Xataka | Agentic AI was the new race for Big Tech and Meta was far behind. It has bought the company most capable of recovering

What is it, how does it work and how safe is this application to install OpenClaw on your mobile with just a couple of clicks

by

Let’s explain to you What is QuickClaw and how does it work?an application created by developer Max Blade to install OpenClaw easily on your mobile. Thus, you can have this popular AI agent without having to deal with APIs, servers and long configuration processes. The application does all this in a completely private way, and creating the agent within a virtual machine for your greater security. You just have to go to their website quickclaw.appclick on the download that is currently exclusive for iOS, and install everything with a single touch. What is QuickClaw QuickClaw is a mobile application that allows you to use OpenClaw easily on your mobile. Install OpenClaw It is a fairly complex process that can put many people off, and what this does is allow you to install it with a single click and without the need for complications or technical knowledge. Thus, you can have your own agent artificial intelligence on your mobile. This means that you will have an AI that does not limit itself to responding to what you ask or perform simple tasks, but can take control of your applications and perform tasks in multiple steps. All of this is done relatively safely. I say relatively because it will use an isolated workspace within your mobile, although the application will also needs access to a lot of your data to function, which may compromise your privacy. In exchange for this, you will have an application that will be able perform various types of tasks for yousuch as creating or editing files, browsing the Internet and taking actions there, managing your calendars, managing your reminders and more. How QuickClaw works The first thing you have to do is install the QuickClaw application, which at the moment It is only available for iOS in the App Store from Apple. When you install the app, you just have to launch it and An OpenClaw instance will be created and launched on a virtual machine “safe”, so as not to touch your device or its operating system. All this in just one minute. The process will then create a Gateway or user service that remains running, in addition to other components that allow OpenClaw to control applications, send messages and execute actions. All this will be done automatically with the recommended OpenClaw installer. This application is configured so that OpenClaw connect to Claude’s AIwho will be in charge of understanding your requests and carrying out the tasks you ask of her. All of this will be designed to be able to control the agent from your mobile using the app itself. What you can do with this app These are some of the tasks you will be able to perform from your mobile just by asking OpenClaw through QuickClaw: Write documents, essays and code, which are delivered as real files. Browse the web and research topics on your behalf. Set reminders and wake-up calls with personalized summaries. Read and manage your calendar and emails. Create, edit and organize files in your private workspace. Execute multi-step tasks autonomously. Remember the context of the conversations. How secure and private is the app One of the features of QuickClaw is to isolate OpenClaw in a virtual machine. This is an important security measureso that nothing the AI ​​agent does can damage the operating system of your mobile device. OpenClaw is an open source agent, which makes it transparent and secure, people can check that it doesn’t do more than what it says it does. However, QuickClaw is not open sourceso you have to trust the word of its creator. Lastly, you should know that everything you make is sent to Anthropicthe company behind Claude. Currently this company is known for its commitment to offering ethical AI, although this can always change in the future. Your data within QuickClaw will be encrypted and anonymous. In addition, your conversations stay within your mobile, since everything runs locally. In Xataka Basics | Dangers of OpenClaw (formerly Moltbot or Clawdbot) and how to protect yourself from them before deciding to use it

OpenClaw is the total AI agent that challenged Big Tech. Big Tech’s response: buy it, of course

by

Peter Steinberger It was a great unknown to the vast majority of the planet until less than a month ago. His project, which he initially called Clawdbot (later Moltbot and finally OpenClaw), became the new sensation of the internet and the world of AI. Its growth has been so spectacular that the majors in this segment set their eyes on it and, inevitably, began to fight to sign its creator and acquire his project. We already have a winner of that bid: OpenAI. What is OpenClaw. OpenClaw is what we could define as “the total AI agent.” A system that uses one or more AI models such as those from OpenAI, Anthropic or Google to do things for you. Here are some differences from using those models in a “traditional” way: You can chat with your AI agent using messaging apps like Telegram or WhatsApp, as if it were just another contact OpenClaw takes full control of the machine you install it on, whether it’s an old PC, a Raspberry Pi or a VPS, for example. You have permission to do whatever you want inside that machine, which also involves risks The capacity of current models, such as Opus 4.5, makes the agent certainly autonomous and proactive and, for example, suggests things to you or makes decisions based on the conversations you have with him? she? it? OpenAI buys OpenClaw. Last week Steinberger I already commented in an interview with Lex Fridman that OpenAI and Meta had made offers to sign him and acquire his project. Those intentions crystallized on Saturday, when the creator of OpenClaw advertisement that he had signed with OpenAI and that the OpenClaw project “will become managed by a foundation and will remain open and independent.” It was a more than reasonable exit for Steinberger, who will probably have received a significant sum of money and prestige, but that leads us to the eternal question: can you compete with the big companies? Short answer: probably not. Large companies have always been hampered by their own size when it comes to reacting quickly to new trends, and even the largest AI companies suffer from this same problem. OpenClaw was doing something that none of them had dared to do – partly because this type of agent has too much “power” – but with these projects and with startups that are beginning to emerge, the same thing always happens: either the big companies copy the idea and they end up burying the originalor they buy that startup that threatened to compete with them. For many startups, in fact, the “exit” or future strategy of the project happens to be bought by a large company. A creator who didn’t want to be CEO. Steinberger explained in his post how his project opened up “an endless string of possibilities” for him, and confessed that “yes, I could really see that OpenClaw could have become a giant company. But no, I’m not excited about that. I’m a creator at heart.” Steinberger has already created a company and dedicated 13 years of his life to it, and “what I want is to change the world, not create a big company, and partnering with OpenAI is the fastest way to bring this to the entire world.” One person’s first unicorn? The appearance of ChatGPT soon made will be spoken of the ‘Solo Unicorn’ phenomenon, a startup created by a single person and which, thanks to AI, would be valued at more than 1 billion dollars. We do not know what price OpenAI has paid for this signing, but it is likely that it will not reach that much. What does seem evident is that OpenClaw was the type of project and idea that certainly could have turned it into that “Solo Unicorn”. The era of custom AI agents. Sam Altman, CEO of OpenAI, confirmed the news in X. There it indicated that the creator of OpenClaw had joined OpenAI “to lead the next generation of personal agents”, and highlighted that “we expect this (personalized AI agents) to quickly become an integral part of our product offerings.” In addition, he assured that OpenClaw will remain open source, something that was probably one of the essential conditions that Steinberger set to join the ranks of OpenAI. And now what. That the project remains Open Source and independent is great news and theoretically that will allow OpenClaw to continue functioning as before, but having OpenAI’s resources can undoubtedly make it grow exceptionally. It remains to be seen whether that will end up having a negative impact in any way, but what also seems clear is that these types of “full AI agents” could soon also be an integral part of the offering of other AI companies. Welcome to the era of total AI agents. We had already partially seen what OpenClaw does with projects like Computer Use from Anthropic, Project Jarvis/Mariner by DeepM Mind u Operator from OpenAI itself. Both allowed AI would do things for us in the browser, but OpenClaw does things for us with all the applications on the machine on which we install it (the email client, the command console, etc.). We are facing an interesting stage for this type of systems. In Xataka | OpenClaw is one of the most fascinating and “dangerous” AIs of the moment. A Malaga company has come to the rescue

How to uninstall OpenClaw (formerly Moltbot or Clawdbot) safely on your computer step by step

by

Let’s tell you how to securely delete OpenClawformerly known as Moltbot or Clawdbot. This is a AI agentwhich can control your computer entirely, and which you can then control through apps like WhatsApp or Telegram. But there are quite a few dangers of OpenClawso just as we teach you how to install it If you want to poke around, now we will tell you how to uninstall it. As with installation, uninstallation requires a series of commands in the terminal from your computer, be it PowerShell in Windows or the terminal in macOS and GNU/Linux. 1. Use the official uninstall command The first step is to open your computer terminal and paste the following command*. When you do this, it will delete the local database, workspace, and installation files. openclaw uninstall –all –yes In the command, you have to use the name with which you installed the agentwhich can be openclaw, clawdbot or moltbot. In the first few weeks of this AI, the name has changed several times. It is possible that when finished you will see a message telling you that CLI is still installed. This is normal, and is something we will fix in the third step. 2. Manual cleaning If you use Windowspress Win+R and write services.msc. Within the screen where you go, look for the ClawdBot, MoltBot or OpenClaw service, right click and choose the option to stop it. Now, you have to open file explorer and delete these folders if they exist. The folder that should appear will depend on the name with which you installed the tool. C:\Users\\.clawdbot C:\Users\\.molbot C:\Users\\.openclaw If you use macOSenter the terminal and type one of the following commands, depending on the name with which you installed them. rm -rf ~/.clawdbot rm -rf ~/.moltbot rm -rf ~/.openclaw Then, you have to type one of these commands to prevent Clawdbot from starting every time you turn on your computer. rm -rf ~/Library/LaunchAgents/com.clawdbot rm -rf ~/Library/LaunchAgents/com.moltbot rm -rf ~/Library/LaunchAgents/com.openclaw 3. Remove the CLI tool You have already deleted your data, but the tool and its commands still exist in your terminal, since you installed them through NPM or HomeBrew. So, now you will have to uninstall it. If you have installed the tool through NPMwhich is the most common, use one of these commands, depending on the name you installed the tool with: npm uninstall -g clawdbot npm uninstall -g openclaw npm uninstall -g moltbot If you are a Mac useryou may have installed it through Homebrew. In this case, if the commands above do not work for you, you can try these: brew uninstall clawdbot brew uninstall openclaw brew uninstall moltbot If none of the commands above work for youthere is another more manual method if you don’t remember how they were installed. They are these commands: sudo rm -f /usr/local/bin/clawdbot sudo rm -f /usr/local/bin/openclaw sudo rm -f /opt/homebrew/bin/clawdbot 4. Additional security cleaning Ok, we have already removed the agent from our computer, but still has access to things like your email, your calendar, or your cloud accounts if you gave it to them when you installed it. Now it will be time to clean these permissions. To do this, you have to review the API keys you have collectedsuch as GitHub, Google Cloud, AWS, etc. Review and revoke or regenerate any API you gave to the bot. You also have to disconnect applications where you have linked it, such as your Google or Microsoft accounts, to revoke access to OpenClaw. Finally, log out of the agent web tool to ensure that you have stopped using it completely. 5. One last check Once you have done all this, restart the computer and open the terminal to write the following code: clawdbot –version If the response message is that it has not found the command or does not recognize it, it is because that’s it, you are done with the uninstallation. In Xataka Basics | The best prompts to save hours of work and do your tasks with ChatGPT, Gemini, Copilot or other artificial intelligence

OpenClaw is the most viral, fascinating and dangerous AI of the moment. For this last reason, it has joined forces with VirusTotal from Malaga

by

In 2025 we had a ‘DeepSeek moment’ and in 2026 we are having an ‘OpenClaw moment’. This AI agent is super powerful, but also super insecure. There is, however, good news, because the Malaga company VirusTotal has partnered with the OpenClaw project to try to mitigate one of the most important cybersecurity risks of this AI agent: its skills. what has happened. OpenClaw (formerly Moltbot, and before Clawdbot) has announced that it has begun a collaboration with the Malaga cybersecurity company VirusTotal, owned by Google. The agreement will see VirusTotal be in charge of “scanning” and analyzing the so-called “skills”, which work like OpenClaw plugins and add all kinds of functions. They do it, of course, but many take the opportunity to introduce malicious instructions that allow them to steal data and remotely operate other people’s AI agents. More security for disturbing AI. Peter Steinberger, creator of the project, has joined Jamieson O’Reilly, cybersecurity expert and founder of the company Dvulnand Bernardo Quintero, founder of VirusTotal, to offer that “additional layer of security for the OpenClaw community.” In it official announcement explain that “all the skills published in ClawdHub (the project’s official skills “store”) are now scanned through Virus Total’s Threat Intelligence system, including its new capability Code Insight (code inspection)”. Bernardo Quintero indicated on Twitter how the effort has already allowed 1,700 skillls to be identified as malicious. If the skill is malicious, it is blocked. This analysis carried out with the VirusTotal tools allows us to identify skills as malicious and block them immediately so that they cannot be downloaded. Not only that: those skills that have been classified as benign are analyzed again every day to detect scenarios in which for some reason they could end up becoming malicious. Still, be careful. Those responsible for OpenClaw warn: the VirusTotal scan helps a lot, but it is not a total guarantee that any skill can perform malicious actions on the machine on which we have our AI agent installed. The attacks of prompt injection Sophisticated skills can manage to cross that barrier, but of course this collaboration means that OpenClaw users can be much calmer regarding the skills available in the ClawdHub repository. OpenClaw wants to be much more secure. This first effort joins OpenClaw’s ambition to have a complete cybersecurity model which includes things like a public roadmap for your new developments in this area, a formal communication process, and details about full audits of your code. Plugging a problem that could kill OpenClaw. The OpenClaw project soon went viral due to its eye-catching options, but shortly after doing so a security audit initial 2,851 skills detected 341 malicious skills. Companies like BitDefender also joined these efforts to avoid problems with tools like AI Skills Checker to check whether a skill was dangerous or not. These malicious skills were, for example, capable of executing shell commands on the victim machine, which gave the attacker complete control of those resources. Attacking the machine is confusing it with natural language. Normally cybersecurity attacks are complex, but the problem with AI agents is that they work with natural language. This implies that to infiltrate these systems you do not have to use code, but simply “convince” and “trick” the AI ​​with natural language. That is where prompt injection attacks come in, which consist of giving instructions to those AI agents that can confuse them to obtain something that theoretically they should not allow them to obtain. Personal data, API keys of the models we use at OpenClaw, email accounts and passwords for all types of services… the possibilities are endless, and OpenClaw, which has access to all of this to operate autonomously, can end up being “tricked” into transferring said data. Beware of OpenClaw. These problems now seem a little less feasible thanks to the collaboration with VirusTotal, but those who are trying OpenClaw on their machines or any other platform should be very alert from the beginning. There are guides that help you install it with some barriers important security issues, and the project itself has a command (‘openclaw security audit –deep –fix’ to audit the most important problems and address them. In Xataka | OpenAI has a problem: Anthropic is succeeding right where the most money is at stake

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.