External processing is always a separate choice
Local maps never call a model. An AI job is created only after you choose a capability, review the provider disclosure, and confirm that the selected source or prompt may be transmitted to FAL for processing. Provider links can expire, so download useful results promptly rather than treating them as cloud storage.
API keys stay on the Worker. The browser submits an allowlisted capability and its input; the server selects the model, calculates credit cost, reserves credits, and records an idempotency key before contacting the queue.
Choose among five consent-gated AI texture workflows
The AI texture generator exposes five allowlisted capabilities: scene-aware depth estimation, PBR map extraction, material generation from a prompt or reference, photographed material-channel extraction, and seamless image upscaling. This covers text-to-texture and image-to-PBR AI tasks without pretending that every request follows the same model path. Each job has a different unit cost, so the server prices by capability, requested megapixels, and requested map count.
A queued response is not treated as a completed result. The job remains pending until the provider reports success. A provider rejection, timeout, or terminal error releases the reservation and records a concise failure code without exposing a raw secret-bearing response.
Credits keep variable model cost separate
Pro Local is a one-time license for browser-side productivity. AI work is sold separately because each request incurs provider cost. Starter, Creator, and Studio credit packs let occasional and production users choose an appropriate amount without forcing a recurring subscription for the basic generator.
No plan promises unlimited model use. The visible estimate is informational; the server-owned ledger is authoritative. Reusing the same idempotency key returns the existing job instead of reserving a second charge.