When GitHub launched Copilot and began automatically completing lines of code – and later entire code snippets – many people asked the question: How long before we can simply describe an app in natural language and Copilot creates it for us? We've seen some experimentation in this area in recent months, but now GitHub itself is championing the idea with the announcement of GitHub Spark at the company's annual GitHub Universe conference in San Francisco.

Spark, officially an experiment the company is launching in its GitHub Next labs, allows you to quickly build a small web app that uses only natural language. Experienced developers can still see and edit the code – and among them are a GitHub repository, GitHub Actions and Microsoft's Azure CosmosDB as the default database for applications that need one – but that's optional. Ideally, you can create a prototype in a chat-like experience and then refine it in subsequent steps.

To build an app, start with an initial prompt and Spark will show you a live preview within a few seconds. From there, you can tweak and iterate the app by going back and forth with the bot.

“We strongly believe this is the next step in building personal software,” GitHub CEO Thomas Dohmke told me. “These micro-apps are not intended to replace the professional developer. They are intended to be a tool with which you can explore ideas, build little bots and little helpers for your everyday life – or simply explore natural language software development.”

While the GitHub materials mention Spark as a tool for building “micro-apps,” Dohmke told me he wasn't really sure if there were built-in limitations on the complexity of a Spark application, and a GitHub spokesperson told me confirmed that currently there are indeed no limitations on prompts or capacity (although this may change after the technical preview). Dohmke noted that Spark can use any web API and leverage AI models on its own, and that he's excited to see how far GitHub users will push this tool.

“Honestly, it's going to be very exciting to see what users can do, how big of an application they can build with just the power of natural language and this simple interface that's really designed to try out ideas and stimulate new ideas.” he said.

With GitHub Copilot now allowing users to choose which major language model they want to use, it's no surprise that Spark offers the same functionality, with users choosing between the latest versions of Anthropic's Claude Sonnet and OpenAI's GPT models.

Users can easily share their Sparks with custom access controls. Perhaps even more interesting, these users can also use and build on the shared code themselves.

Developers who want to develop these applications even further can take a look at the code at any time – and edit it if there is a problem that Dohmke says might arise. After all, AI is not perfect. “You can of course look at the code base,” he said. “So if you understand the code base, you can also look at the code directly and change it, which is often helpful when the AI ​​makes a mistake – which is what happens.”