Google AI Studio adds multi-chat and web search to Build mode

• GoogleAIStudio's launch post says Build mode now has two new primitives: multi-chat threads and web search, both aimed at keeping coding sessions current without leaving the workspace.
• OfficialLoganK's announcement frames multi-chat as multiple threads per app, so a single project can branch into separate conversations instead of forcing one long context window.
• Search grounding in Build mode mirrors Google's existing Grounding with Google Search docs, which describe Gemini's google_search tool as a way to pull real-time web content and attach citations.
• Google's earlier full-stack vibe coding launch post and AI Studio build docs make clear this feature lands inside AI Studio's browser-based app builder, not as a new Gemini API product surface of its own.

You can open Build mode here, skim Google's vibe coding page, and compare the new search button to the older Gemini grounding docs. The more interesting detail is scope: Google's March writeup pitched AI Studio as a full-stack browser builder with server-side runtime and one-click deployment, while one early user reply already asked for GitHub branch creation and merge flows that still are not in the product itsPaulAi's reply.

Google's wording is specific. According to OfficialLoganK's post, multi-chat means "multiple chat threads per app," so the unit of organization is the project, not the account.

That fits how Google has been positioning AI Studio Build mode since March. In Google's full-stack launch post, the Antigravity agent was already framed as a browser-based app builder that can iterate on an existing project, keep context about the codebase, and publish with one click. Multi-chat adds a cleaner branching mechanic on top of that existing loop.

The second ship is live web retrieval inside the coding session. GoogleAIStudio's announcement says the agent can "pull real-time info from the web to ground your session in the latest documentation," which is a useful upgrade for in-browser prototyping because dependency docs and API references go stale fast.

Google's Grounding with Google Search documentation describes the same underlying capability in more concrete terms: the google_search tool connects Gemini to real-time web content, can reduce hallucinations, and automatically returns citations. What changed here is packaging, that API-era grounding feature is now directly wired into Build mode's UI.

Early reaction surfaced the next obvious missing piece. In itsPaulAi's reply, a user says AI Studio would become their main building tool once it can create a new GitHub branch and merge from inside the product.

That comment lines up with the current product boundary. Google's vibe coding page sells prompt-to-app building, remixing, and one-click launch, and the full-stack docs emphasize server-side runtime, secrets, npm packages, and multiplayer features. Neither page describes native branch management or merge workflows yet.