You found the edge of the demo.
This preview is read-only on purpose — no accounts, no server, nothing to break. The part you just reached for opens with the hosted release.
Opening with the full release
Accounts & saved brands
Sign in, keep every brand, and pick up where you left off — across devices.
Live AI generation
Routed per asset type across Gemini, Veo 3.1, and OpenAI's gpt-image-2 — whichever tested best for the job — with a free deterministic result when no provider can deliver.
Plans & credits
Simple tiers, and credits only where AI does the work — free heuristic tools stay free.
CLI · MCP server · API
The same engine, scriptable: a CLI for your terminal, typed MCP tools for AI agents, API access on Studio and Scale.
Everything above is already built and running locally — the release wires it to the world.
Meanwhile, the demo has the good part: four complete brands generating themselves, every asset the product’s real output.