Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import { useEffect, useState } from 'react'
/**
* Whether a local Evenfire is answering on this machine.
*
* One bounded probe per mount. The main process takes no URL — it only ever
* checks the built-in Localhost option — so the renderer cannot use this to
* reach an arbitrary host. A failed or slow probe resolves to `false`, which
* every caller renders as "show nothing": a machine with no local cluster
* never sees a Localhost affordance advertising a server that isn't there.
*
* Optional-chained throughout because tests and stale dev preloads can leave
* `window.clerum` partially defined, and a missing bridge must degrade to
* "not reachable" rather than throwing during render.
*/
export function useLocalhostReachable(): boolean {
const [reachable, setReachable] = useState(false)
useEffect(() => {
let cancelled = false
const probe = window.clerum?.auth?.probeLocalhostReachable
if (typeof probe !== 'function') return
probe()
.then(result => {
if (!cancelled) setReachable(Boolean(result))
})
.catch(() => undefined)
return () => {
cancelled = true
}
}, [])
return reachable
}