The all-in-one local AI studio for your desktop: chat, image and video generation and a coding agent in one free, open source app. Windows and Linux. No Docker, no terminal, no cloud required.
TypeScript agentai-chatai-studioall-in-onecoding-agentdesktop-appfluximage-generationlinuxllmlocal-ailocal-llmopen-sourceprivacyself-hostedstable-diffusionuncensoreduncensored-aivideo-generationwindows
import { useEffect, useState } from 'react'
import { getContentPolicy, type ContentPolicy } from '../api/cloud/jobs'
/**
* C2: the account's content policy, fetched once and shared across every
* surface that only needs it for DISPLAY (today: ModelChip's "No Filter"
* mark). Mirrors the web app so the same
* rule reads the same way on both sides: a small module-level cache plus a
* listener set, instead of every mounted ModelChip firing its own GET.
*
* Read-only. The server re-checks the real policy on every render job
* (ContentPolicySettings.tsx already owns the write path); a stale or wrong
* value here mislabels a badge, it grants nothing.
*/
let cached: ContentPolicy | null = null
let inflight: Promise<ContentPolicy> | null = null
const listeners = new Set<(p: ContentPolicy) => void>()
export async function loadContentPolicy(): Promise<ContentPolicy> {
if (cached) return cached
if (!inflight) {
inflight = getContentPolicy()
.then((s) => {
// A concurrent write (ContentPolicySettings' save()) may already have
// primed the cache while this GET was in flight, and that newer value
// wins, this response must not overwrite it.
if (cached) return cached
cached = s.policy
listeners.forEach((l) => l(s.policy))
return s.policy
})
.catch(() => 'soft' as ContentPolicy)
.finally(() => { inflight = null })
}
return inflight
}
/**
* The policy as currently known, without triggering a fetch. `null` means
* "not loaded yet"; a caller that needs to act on it (B3: the client-side
* cloud adult gate in useCloudCreate) must treat that as "unknown", not as
* a guessed rule, mirroring the web app's contentPolicySnapshot().
*/
export function contentPolicySnapshot(): ContentPolicy | null {
return cached
}
/** Push a freshly saved policy into the shared cache (call after a
* successful PUT), so every mounted ModelChip picks it up immediately
* instead of waiting for a reload. */
export function primeContentPolicyCache(policy: ContentPolicy): void {
cached = policy
listeners.forEach((l) => l(policy))
}
/** Test-only: drop the cache between cases. */
export function _resetContentPolicyCache(): void {
cached = null
inflight = null
}
export function useContentPolicy(): ContentPolicy {
const [policy, setPolicy] = useState<ContentPolicy>(cached ?? 'soft')
useEffect(() => {
listeners.add(setPolicy)
void loadContentPolicy().then(setPolicy)
return () => { listeners.delete(setPolicy) }
}, [])
return policy
}