After much fanfare, GitHub today officially announced a brand new product called GitHub Spark, along with updates to GitHub Copilot. GitHub Spark is essentially an AI-powered tool that allows anyone to create small web applications (sparks) using natural language, even without any programming or web application development knowledge.
GitHub describes GitHub Spark as follows:
It all starts with an idea. As kids, we have thousands of ideas—some silly, some crazy, some fun. Today, we’re introducing GitHub Spark, a product that’s born from fun ideas. Powered by natural language, it’s the foundation of our vision to help 1 billion people around the world become developers. This is fun, personal software, not an enterprise app.
GitHub Spark features interactive previews, revisions, automatic history, and model selection for development. Available from both desktop and mobile. Spark has three tightly integrated components:
- NL-based editor to easily describe and refine app ideas.
- The managed runtime environment hosts sparks, providing data storage, topics, and access to large language models (LLMs).
- PWA powered dashboard to manage and launch sparks from anywhere.
Sparks created with GitHub Spark can be shared with others with read-only or read-write permissions. Users can choose from a variety of models, including Claude Sonnet 3.5, GPT-4o, o1-preview, and o1-mini.
In the coming months, GitHub will update GitHub Spark in the following aspects:
- Extend collaboration methods (e.g. public libraries, allowing users to perform semantic merges of changes someone made in a branch of spark, multi-player)
- Extend the editor surface (e.g. provide an “x-ray” mode that allows for summarizing and adjusting the exact behaviors of the application)
- Extend the runtime environment (e.g. more built-in components, better integration with 3rd party services, allowing file storage and vector searching).
You can sign up for the GitHub Spark technical preview HERE. With GitHub Spark, creating and sharing small web applications is as easy as typing a simple sentence of text. This innovative tool has the potential to “democratize” software development and empower a new generation of makers.