Agent S

Introducing the Gemini 2.5 Computer Use model, a specialized agent model built on top of Gemini 2.5 Pro’s visual reasoning capabilities, designed to interact directly with user interfaces (UIs). It is exposed via a new computer-use tool in the Gemini API, with inputs that include the user’s request, a screenshot of the UI environment, and a history of recent actions. The model generates function calls corresponding to UI actions like clicking, typing, or selecting, and may request user confirmation for higher-risk tasks. After each action is executed, a new screenshot and URL are fed back into the model to continue the loop until the task completes or is halted. It is optimized primarily for web browser control and shows promise for mobile UI interaction, though it is not yet suited for desktop OS-level control. In benchmarks across web and mobile control tasks, Gemini 2.5 Computer Use outperforms leading alternatives, delivering high accuracy at lower latency.

OpenClaw is an open source autonomous personal AI assistant agent you run on your own computer, server, or VPS that goes beyond just generating text by actually performing real tasks you tell it to do in natural language through familiar chat platforms like WhatsApp, Telegram, Discord, Slack, and others. It connects to external large language models and services while prioritizing local-first execution and data control on your infrastructure so the agent can clear your inbox, send emails, manage your calendar, check you in for flights, interact with files, run scripts, and automate everyday workflows without needing predefined triggers or cloud-hosted assistants; it maintains persistent memory (remembering context across sessions) and can run continuously to proactively coordinate tasks and reminders. It supports integrations with messaging apps and community-built “skills,” letting users extend its capabilities and route different agents or tools through isolated workspaces.