Opal is now inside the Gemini web app, which means you can build reusable AI mini-apps right where you already manage Gems. If you’ve been waiting for an easier way to create custom Gemini tools without writing code, this is Google’s latest experiment to try.
Google Labs describes Opal as a visual, natural-language builder for multi-step workflows, the kind that chain prompts, model calls, and tools into a single mini app. Google also says Opal handles hosting, so once an app’s ready, you can share it without setting up servers or deploying anything yourself.
Go to Gems manager and try it now
To find Opal in the Gemini web app, head to your Gems manager and look for the Opal option to create experimental Gems. From there, you can type what you want the mini app to do in plain language, then edit the generated workflow and reuse it later for the same task.
Google’s pitch is speed-to-prototype. Instead of rebuilding the same prompt chain every time, you can turn it into something you can run again and again.
A new step list view
Opal’s editor is also getting a clearer way to understand what you’ve built. Google says you’ll see a view that converts your prompt into a step-by-step list, so it’s easier to spot what each part does and where a change should go.
That’s a practical upgrade because “no code” only works if editing stays approachable. If a single step is off, you can tweak that step instead of rewriting the whole thing.
If you need more control, Google points builders to the Advanced Editor at opal.google for finer-grained edits. Think of Gemini’s Opal entry point as the quick build path, and the Advanced Editor as the place you go when workflows get complex.
Both are available today: Opal in the Gemini web app via Gems manager and the standalone Advanced Editor at opal.google. Try it out now and compare it to the best AI tools out today.