An artificial intelligence technology developed by Google LLC lets people create basic apps without knowing any coding.
Opal, the tool, was introduced on Thursday. The debut follows a few weeks after Amazon Web Services Inc. unveiled Kiro, an IDE (integrated development environment) that leverages artificial intelligence to expedite software development.
An Opal development project begins with a chatbot interface similar to ChatGPT. Users describe the task their application will complete in natural language. They can also connect external tools and supply any further information that could be required to finish the task.
The service automatically creates a basic application using the supplied data. A button in the upper right corner of the UI allows users to distribute the software once it is ready. It makes the application available via a sharing link that resembles the ones that Google Docs uses.
Employees have the option to use a visual editor to make changes to an Opal-generated application before publishing it. An application is represented by the editor as a set of cards on a digital canvas. While some cards regulate how the application processes the input it receives, others describe the input it takes in.
Multiple data processing processes can be linked together thanks to the visual editor. A user might, for instance, instruct an application to provide a description of a product and then utilize that description to produce a promotional video. By using Opal’s visual editor to enter natural language instructions, employees can personalize how these activities are completed.
In a demonstration video, Google product manager Elle Zadina described how Opal transforms your app description into a multi-step workflow with inputs, generation steps, and output steps. “You can edit the prompt or instructions directly by clicking into a step.”
Opal uses a variety of AI algorithms to create apps under the hood. Google did not identify the algorithms that are employed. The company’s recently released Gemini 2.5 Pro large language model is one contender; it excels at tasks involving application development. On the well-known LMArena benchmark for evaluating LLMs’ coding proficiency, it has achieved a new record.
Opal might not need Gemini 2.5 Pro’s sophisticated programming features because it is designed for developing basic apps. It’s possible that Google is powering the tool with a less powerful but more power-efficient coding model. In turn, one of Opal’s Veo video algorithms is most likely the basis for the clip generating functionality.
Over six pre-built applications are being released by Opal. Some concentrate on creating video games and other consumer use cases. The remainder allow employees to conduct product research, create marketing collateral, and carry out other business-related duties.
Opal is first accessible to U.S. users via a public beta test. Google Labs, a website that houses over two dozen of the search engine giant’s experimental AI technologies, is where users may use the tool. They automate everything from web searches to designing application interfaces.