Manus, the AI startup recently acquired by Meta, has unveiled a significant evolution in its artificial intelligence agent technology with the launch of Manus Desktop. This new application brings its powerful, multi-step task execution capabilities directly to personal laptops, marking a strategic shift from its previous cloud-exclusive model.
The ‘My Computer’ feature within Manus Desktop empowers its AI agents to interact seamlessly with local files, applications, and tools residing on a user’s device. This development positions Manus, and by extension Meta’s AI efforts, in closer competition with emerging open-source AI agents like OpenClaw.
OpenClaw, founded by Austrian developer Peter Steinberger, has rapidly gained traction since its introduction late last year, fueling a burgeoning interest in AI agents capable of local execution. The trend has captured the attention of industry leaders, with Nvidia CEO Jensen Huang notably describing OpenClaw as “the next ChatGPT” in a recent television appearance. Steinberger’s subsequent hiring by OpenAI, a key AI competitor to Meta, further underscores the strategic importance of this evolving AI landscape.
Unlike the open-source and free-to-use nature of OpenClaw, Manus primarily operates on a paid subscription model. However, the ‘My Computer’ functionality promises substantial utility, enabling its agents to perform tasks such as reading, analyzing, and editing files, as well as launching and controlling applications. Manus highlights potential use cases like organizing vast local image libraries and even accelerating application development, potentially creating new applications within minutes. These capabilities will be integrated with Manus’s existing suite of integrations, which include popular services like Google Calendar, Gmail, and a range of third-party platforms.
This enhanced local processing power, while offering immense convenience and efficiency, also brings to the forefront critical discussions around data security and user privacy. Experts have voiced concerns about granting AI agents direct access to personal devices, citing potential vulnerabilities. Manus addresses these concerns by emphasizing user control, stating that explicit approval will be required before any task is executed. Users will have options to grant permissions on a case-by-case basis (“Allow Once”) or for recurring, trusted actions (“Always Allow”).
Meta’s acquisition of Manus, announced in late 2025, was a clear signal of its intent to bolster its AI capabilities and embed autonomous agent technology across its product ecosystem, including the Meta AI assistant. Manus, originally founded in China before relocating its headquarters to Singapore, has been subject to regulatory scrutiny. Reports indicate that Chinese officials have been investigating the approximately $2 billion acquisition for potential technology control violations.
A Meta spokesperson reiterated to CNBC that the acquisition “complied fully with applicable law” and expressed confidence in “an appropriate resolution to the inquiry.” The company further stated that the Manus team is now “deeply integrated into Meta, running, improving and growing the Manus service and will continue to make it available to the millions of people who enjoy it.” The integration of Manus’s advanced AI agent technology into Meta’s platforms signifies a significant step in the company’s pursuit of more intelligent and capable AI assistants.
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/19858.html
