There's an ongoing refactor/modernization process taking place for SolidStart, that's unfolding throughout 2025. It'll make the framework more maintainable, lean and robust. The motivations are:
- There's a common understanding that the layer of the stack that sits between solid-start and vite+nitro (which vinxi has been handling) has to move forward, also on the shorter term.
- The design choices for vinxi were made ~1.5 year ago, and things have changed a lot since, meaning that vinxi today is significantly more complex than it has to be. Two notable examples here are:
- Vite 6 Environment API. For Vinxi to be as flexible as it is, it needed an abstraction layer over vite, which in practice spawns multiple vite instances. This idea has since been refined an upstreamed as the Vite 6 Environment API, allowing creation of multiple environments in a single vite instance.
- Nitro 3 Vite integration. Vinxi practically showed how Nitro could be used with Vite in a generic way, and to do this Vinxi served as a bridge to Nitro. This ideas has since been refined and upstreamed in the nascent Nitro 3, which will have better options for direct Vite integration.
As a result, this modernization process is now ongoing, with an end-goal to reimagine/replace vinxi. This will happen in practice through the creation of a set of new leaner modules, rather than changes to the vinxi repo. Some of these modules will leverage the new APIs from vite and nitro that vinxi over the past year helped motivate. It's expected to increased the maintainability and reliability of this layer in the stack.
Since TanStack and SolidStart both have been powered by Vinxi up till this point, this transition happen as a collaborative effort.
Progress Tracker