It is based on the lumaAI official response, which requires queuing and sometimes not so fast to generate videos