GitHub's new tool helps you build apps using plain English
GitHub has unveiled a new experimental tool, Spark, at its annual GitHub Universe conference. The innovative platform lets you create small web apps with natural language. The project, which is part of GitHub Next labs, promises a chat-like experience for app development. It starts with an initial prompt from which Spark generates a live preview in seconds, letting you refine your app by interacting with the bot.
A tool for personal software creation
GitHub CEO Thomas Dohmke sees Spark as a major step in personal software creation. He explained that these "micro apps" aren't meant to replace professional developers, but are more like tools for exploring ideas and building small bots or helpers. Despite being touted as a tool for "micro apps," there are no limits on the complexity of apps you can build with Spark, a GitHub spokesperson told TechCrunch.
Capabilities and user customization
Spark can use any web API and add AI models. Dohmke said he was excited to see how users will use this tool. "It will be very exciting to see what users can do, how big of an application they can build with just the power of natural language," he said. Like GitHub Copilot, Spark also lets users choose their preferred large language model from latest versions of Anthropic's Claude Sonnet and OpenAI's GPT models.
Just give the command to get the result
Spark allows code sharing and modification
Spark also lets users share their Sparks with customized access controls, allowing others to build on shared code. However, Dohmke admitted that AI isn't perfect and there may be cases where developers would have to tweak the code. He said, "You can, of course, look into the code base... So if you have an understanding of the code base, you can also look at the code directly and modify that."
