On May 7, the tech media outlet Analytics India Magazine reported that Hugging Face has launched a new cloud-based AI tool called Open Computer Agent. This innovative tool allows users to remotely control a Linux-based virtual machine using simple text commands, including the ability to run applications like Firefox.
The Open Computer Agent integrates technologies such as SmolAgents, the Qwen2-VL-72B visual language model, and E2B Desktop, enabling users to execute straightforward commands like launching applications.
Equipped with commonly used applications like the Firefox browser, the tool effectively responds to basic English commands for tasks like opening websites or searching for directions. However, during early tests, users noted issues with response speed and performance stability, particularly when handling more complex tasks. Instances of errors, especially during CAPTCHA verification processes, were reported.
While the tool is now publicly available, high demand means users may face a wait in a virtual queue for access. Hugging Face clarified that the aim of this tool is not to achieve perfection but rather to demonstrate the competitive edge and cost-effectiveness of running open-source models in the cloud.
