Google Enhances Opal, Its AI ‘Vibe-Coder’

Google updates Opal AI tool with new automated workflow creation method.

Google has updated its AI tool for vibe-coding, Opal. A new method for creating automated workflows has been added.

Opal, our no-code visual builder for AI workflows, just got a major upgrade. 🧠💎

We’ve added a new agent step that analyzes your goal, determines the best approach, and automatically calls the right tools — such as Veo for video or web search for research — to complete the… pic.twitter.com/QnKlG0b4ZJ

— Google Labs (@GoogleLabs) February 24, 2026

The agent allows for the creation of mini-applications for planning and executing tasks through text prompts. The feature is based on the Gemini 3 Flash model, which automatically selects the necessary tools.

Potential scenarios include:

• integration with Google Sheets to save data between sessions (for example,
• using web search for information gathering or the Veo model for generating video content.

The assistant independently forms a plan and determines the next step.

Google emphasized that the agents are interactive: if they require additional information, they will request it from the user or suggest their own options to determine further actions.

This enables even users without technical knowledge to build complex workflows in their applications.

Opal was first introduced in July 2025. The tool allows for the creation of mini-applications or the enhancement of existing ones. In December, the service appeared in the web version of Gemini, enabling the assembly of ready-made products in a visual editor without writing code.

Recently, many companies and startups have been focusing on vibe-coding, offering their solutions. Among the most popular are Lovable and Replit, as well as projects like Wabi, Emergent, and Rocket.new.

Major players in the AI industry are also developing this area: OpenAI has Codex, and solutions from Anthropic are in high demand.

Earlier in February, the flagship model Claude Opus 4.6 received a “team of agents.”