Jakarta, INTI - Google announced on Tuesday that it is introducing a new capability for building automated workflows within its vibe-coding application, Opal. According to the company, a newly launched agent inside Opal enables users to develop mini-apps that can plan and carry out tasks through text-based prompts.
The feature runs on the Gemini 3 Flash model and can automatically select the appropriate tools to complete tasks. For example, it may leverage Google Sheets to retain memory between sessions, such as storing a shopping list for an e-commerce app. The agent is also capable of independently determining and organizing the next steps required to complete assigned tasks.
Google explained that these agents are inherently interactive. If additional input is needed, they will prompt users to provide more details or present selectable options to guide the next action. With this update, the company says even users without technical expertise can design sophisticated workflows within their applications.
Opal’s Global Expansion and Growing Competition
Opal was initially launched for users in the United States in July 2025, allowing individuals to create mini web apps or remix existing ones. In October 2025, the platform expanded to 15 additional countries, including Canada, India, Japan, South Korea, Vietnam, Indonesia, Brazil, and Singapore. By November, Opal became available in more than 160 countries worldwide. In December, Google integrated the tool into the Gemini web app, enabling users to build customized applications through a visual editor without the need for coding.
Outside of Google, numerous startups are also developing platforms that allow users to create applications using natural language prompts. Among the most recognized are Lovable and Replit, while emerging players such as Wabi, founded by a former Replika founder, SoftBank and Lightspeed-backed Emergent, and Accel-backed Rocket.new are also attracting growing attention.
Conclusion
Google’s latest Opal update signals a broader shift toward AI-powered app creation, lowering technical barriers for workflow automation. As competition intensifies in the natural language app-building space, intelligent agents capable of autonomous planning may become a defining feature of next-generation development tools.
Read more: Indonesia Partners with ARM to Train 15,000 Chip Design Engineer