Side-questing
backwards

Could an Astro site use Kumo without a React rewrite? I built a small solo experiment to learn. It is unsupported and still in progress. A demo is not a release.

THE QUESTION · KUMO + ASTRO The first long run made output, not a support model. That is useful because it shows the questions we still need to answer. ↗

kumo-compiler.coey.dev · solo exploration

// React Kumo remains the starting point. This experiment does not change it.

A useful exploration has to answer more than “can it render?” It has to leave an honest path for someone else to use it, own it, or decline it.

I dogfood our primitives while I explore. The long run leaves problems behind. I use those problems to make the next long run better.

// A long run is worth more when its problems become the next check, tool, or receipt. Models learn from exhaust too: prompts, tools, and especially corrections. The Reverse Information Paradox · July 12, 2026

The useful result today is not a release. It is a small Astro proof and a clear list of what has to be true before anyone should depend on it.

01A repeatable sync pathReact changes have a known way through.

02Known operating costIt is fast and cheap enough to be practical.

03A willing ownerSomeone can take it over and it keeps working.

Kumo Astro explorationsolo, unsupported, work in progress→ Corrections to gatesshared mistakes turned into local checks→

kumo-astro-demo.coey.dev · work in progress

// stacking tokens includes dogfooding: I use my own tokens to build on our primitives. The point is to combine useful pieces into something new, not build another version of the same thing. The receipts make the next decision cheaper.