Google Opens Gemini and Search to Connected Apps, Letting Assistants Act on Third-Party Services
Google has launched a Connected Apps Platform that lets businesses link their applications to its AI experiences in Search and Gemini, according to TechCrunch, which reported the launch on October 28, 2026.
The platform is in beta and public preview for U.S. developers, who sign up for a waitlist. Developers integrate through a cloud-hosted MCP server (Model Context Protocol, an open standard that lets AI models call outside tools) or Android MCP, using AppFunctions APIs. They then register, test and submit their app through the Google Connected Console. Each developer chooses whether the app appears in Search, in the Gemini app, or in both.
Connected apps let users link third-party services and take actions from inside the assistant. Google's examples include designing a flyer through Canva and building a playlist through YouTube Music. The feature was first introduced in Gemini and Google Search earlier in 2026. In July, Google said users could connect apps directly to AI Mode in Search, citing Instacart, Canva and YouTube as examples.
Google says it has more than 50 connected apps available in Gemini. Partners include Adobe, Experian, Localiza and Peloton. Google's own services are on the list too, among them Workplace, Photos, YouTube Music and Google Home. In Search, connected apps currently include Instacart, Canva, YouTube Music and Yahoo Fantasy Football. Gemini connected apps will come to Search over the next few weeks. Meditation app Calm is set to launch on both Gemini and Search as a new Connected Apps partner.
Once an app is live, developers can track traffic, latency (the time a request takes to complete) and error rates. They can also manage brand identity, including verified logos and authenticated accounts. For engineering teams, a third-party action inside an AI answer becomes a production dependency with its own service metrics, rather than a black-box integration.
The launch builds on a run of agent-oriented announcements from Google this year. At I/O in May, Google introduced Gemini Spark, an AI agent (software that carries out multi-step tasks on a user's behalf) that pulls information from various apps to draft reports and manage schedules, as Reuters reported. Google's Search blog that month described bringing advanced model capabilities to Search, with features that let users invoke agents by asking a question. On October 8, Google Cloud introduced the Gemini agent, a single AI agent for work that can answer questions and handle tasks.
In my view, the more consequential change is on the distribution side. For two decades, reaching a customer meant an app store listing, a web landing page or a search ranking. A connected app now has to be reliably callable by an assistant, through a protocol and an API surface the developer controls. Discoverability depends on whether an agent can complete a task against your service without friction, and the latency and error-rate dashboards make that performance visible enough to act on.
The broader context here is that the request still begins as a plain question in a search box or chat window. A user who asks for groceries to be ordered, a flyer drafted or a playlist built no longer needs to open each app separately. Whether that holds up at scale will depend on how well the integrations perform under real load, and on how clearly brand, account and permission boundaries are handled as more services plug in.
