AI Studio
Not enabled yet — on purpose
AI generation is an external dependency we refuse to fake. The unified API, provider adapter, task model, quota checks and audit trail are built; the generation switch stays off until a provider completes contract, quality, cost, security and rollback acceptance. The endpoints answer HTTP 503 with the stable code AI_PROVIDER_NOT_CONFIGURED — no tasks, no quota, no pretend images.
Live API status
Checking…What's planned
Text-to-image
Describe it, get a wallpaper. Free accounts: 2/day, Pro: 10/day, Enterprise shared: 20/day (UTC reset).
Image-to-image
Restyle an existing image with strength control and content safety checks.
Brand compose
Enterprise Logo and Slogan layers with permission control and crisp export.
Asset intelligence
Optional AI tagging for the library, under its own switch — manual import never depends on it.
Failure rules, up front
Once enabled: one automatic retry, then the quota hold is released and you're notified. Prompts, inputs, outputs, model, cost, latency and review results stay traceable.