Gemini's Computer Use tool goes GA for production browser agents
Google moves screenshot-driven UI control out of preview, ships a zero-dependency API CLI for agents, and now logs new paid requests by default.
Copy markdown
Computer Use is production-ready
Google promoted its Computer Use tool out of preview to generally available in both the Interactions and GenerateContent APIs. You can ship agents that drive real browsers, desktops, and mobile by reading screenshots and returning UI actions — clicks, typing, scrolling, drag-and-drop, hotkeys and navigation (mobile adds open-app and long-press). GA means it's backed for production instead of the preview “may break” caveat, with prompt-injection detection on Gemini 3.5 Flash-Lite and newer.
A zero-dependency Gemini CLI for agents
The new gemini-api CLI entered beta on Oct 27: a stateless, zero-dependency command-line client aimed at coding agents, scripts and CI pipelines. It talks to the Interactions API to generate text and media, analyze files, and run managed agents — a clean way to wire Gemini into a pipeline without pulling in an SDK.
New paid requests are now stored by default
Since Oct 26, new paid-tier projects calling generateContent, streamGenerateContent and batchGenerateContent store requests by default (store=true). Existing projects keep their current setting, but on any new project you now have to opt out — set store=false per request, or turn off Generate Content logging in AI Studio — if you don't want prompts and responses retained.
OpenAI's dot agents can now text you
In a US-only Pro beta, OpenAI’s “dot” agents can now message you over iMessage and RCS, including proactive updates. It turns an always-on personal agent into something that reaches you in your native messages thread — no app open required.