Kmemo

A Ktor server plugin that makes a SemanticCache available to every route handler.

kmemo is coroutine-native, so there is nothing to bridge — this plugin just installs your cache where handlers can reach it, and adds a getOrPut convenience so a handler caches an LLM call in one line, guards and all.

install(Kmemo) { cache = SemanticCache(embedder) }

routing {
    post("/chat") {
        val prompt = call.receiveText()
        val answer = call.getOrPut(prompt) { llm.complete(it) }
        call.respondText(answer)
    }
}