GitHub Spark alternatives for personal apps
If you used Spark for small personal tools, choose your next step around what you still have: an exported project, a deployed app, or an idea you want to rebuild. Kaddy is an option for rebuilding the routine in a shared app home. It is not a Spark import or migration service.
What changed with Spark
GitHub's August 4, 2026 announcement stopped new users and app creation that day, with access for exporting existing apps through August 31. It says deployed apps continue to work after retirement. Separately, the GitHub Models service used by Spark's llm() retired July 30, so those calls no longer work.
That distinction matters: a deployed screen remaining available doesn't mean its AI features still work. The announcement's export deadline has passed. Don't assume an old how-to article's workbench instructions are still available.
If you have an exported project
Keep that copy safe before doing anything else. GitHub points builders toward Copilot in their development environments. Continuing with the existing code may preserve more of your work than recreating the app in another builder.
Review what the project depends on, where its data lives, and which AI calls need replacement. A new host won't automatically supply a retired inference service. An app's source code is also not necessarily a backup of its saved records.
If what you want back is the routine
For a small reading list or meal planner, it may be reasonable to rebuild the useful parts. Write down the fields, the actions you used, and the screen you returned to. Leave behind features you never touched.
Kaddy fits that choice when you want several such tools together, each with its own assistant and memory. There is no direct Spark project import. Recreating the interface doesn't move the old data, and we don't promise automatic conversion.
Use the first-app guide to turn that list into a small brief. Test with fictional records before deciding how to handle your real information.
When another option fits better
If the app is becoming a standalone web product, consider the project approach described in our Lovable and Replit comparison. If local project control is the priority, read local vs hosted builders.
We'd avoid choosing purely on which product looks most like Spark. Write down what you want to preserve: the code, the data, the AI behavior, or simply a useful routine. Those are different jobs, and no alternative should imply it handles all four without checking.