Google is transforming Gemini from a general chat interface into an AI application development platform accessible to everyone. Recently, the company officially integrated the "vibe-coding" tool Opal into the Gemini web app (gemini.google.com), allowing users to now build personalized mini-applications through natural language descriptions directly in the browser, without any coding, and save them as exclusive AI assistants - Gems.

Gems is Google's customized AI assistant system launched in 2024, supporting users in creating tailored intelligents for specific scenarios such as study tutoring, career planning, programming collaboration, and copy editing. The addition of Opal completely changed the way Gems are created: no programming background is needed, and development can be completed just by chatting.

image.png

Users simply need to describe their needs in Gemini, for example, "Help me create an assistant that automatically organizes meeting minutes into to-do lists," and Opal will generate the corresponding application logic and present it in a visual flowchart. Users can drag, rearrange, and connect various processing steps in the interface to intuitively adjust the application's behavior. Even more intelligent is the new "Text to Steps" view added by Opal — which converts users' natural language descriptions into structured operation lists in real time, simplifying the setup process and helping users understand the internal workings of the AI.

For advanced users with higher customization needs, Google provides an entry point to the advanced editor (opal.google.com), supporting more complex logic arrangement and integration with external tools. All mini-applications created through Opal can be saved, reused, and used as Gems in subsequent conversations, making frequent workflows truly become a personal AI toolset.

The launch of Opal marks Google's official entry into the "vibe-coding" trend — a new development paradigm driven by startups like Lovable, Cursor, and Wabi, which is making application building shift from "writing code" to "expressing ideas." With Gemini's large user base and multimodal model capabilities, Google has chosen a triple combination of graphical interface + conversation-driven + Gems ecosystem to lower the barrier to AI application creation, attracting non-technical users to participate in creation.