Google expands Opal no-code AI builder to 15 countries

Two months ago, Google Labs quietly rolled out Opal to test whether regular users could build mini-apps just by chatting with an AI. The experiment worked a little too well.

Instead of just churning out silly novelty toys, early testers in the U.S. started piecing together surprisingly complex, practical tools. In response, Google just kicked off an international expansion, bringing Opal to 15 new countries including Canada, India, Japan, South Korea, Singapore, Vietnam, Indonesia, Brazil, and several Latin American markets.

Visual building blocks and zero-code debugging

If you haven’t seen Opal in action, it basically brings vibe-coding to people who never want to open a code editor. You describe what you want, and Opal taps various Google models to generate a mini web app.

Once the app is generated, you get a clean visual canvas that lays out every input, output, and generation step. You can click on any node to tweak the underlying prompt, or manually drag in new steps. When it looks right, you just generate a web link and share it with anyone using a Google account.

Along with the geographic rollout, Google pushed out some much-needed under-the-hood fixes. According to Google Labs senior product manager Megan Li, workflow reliability was one of the biggest community asks.

Debugging now happens step-by-step directly in the visual editor or console panel. Instead of staring at a generic error message, Opal flags the exact node where things broke in real time. Google also enabled parallel execution for multi-step workflows and slashed the initial generation delay, which previously kept users waiting for five seconds or more.

Why it matters

For non-technical creators, marketers, and small teams, Opal is an incredible shortcut. You do not need to beg an engineering team to build a simple text summarizer, quiz generator, or lead-sorting widget. You can build it in Opal during a coffee break, test it immediately, and share the URL across your team.

For the broader AI space, Google is staking its claim in a crowded field alongside Figma, Canva, and Replit. Every tech giant wants to own the canvas where everyday users interact with AI. By giving users a visual pipeline instead of raw code, Google makes AI feel tangible rather than intimidating.

We think Google’s decision to keep debugging strictly no-code is brilliant. Most AI coding assistants force you back into syntax when things break, which defeats the purpose for non-coders. Pinpointing failures to a visual card keeps the experience accessible.

That said, Opal has clear limits. These are lightweight mini-apps, not enterprise-ready software. You will not be wiring up complex transactional databases or handling strict enterprise compliance here. It is a fantastic playground for prototyping, but it is still fundamentally a sandbox.