# Hotel Lobby AI active model and workflow Canonical: https://hotellobbyai.tv/models.md Workspace: https://hotellobbyai.tv/ai-video-generator Pricing: https://hotellobbyai.tv/pricing.md ## MiniMax H3 Public model identifier: minimax-h3 This is the model selected by the effective server-side Hotel Lobby provider configuration, rather than all models left in the inherited template catalog. Provider credentials and internal endpoints are not public model capabilities. - Workflow: two-character reference-video casting in the orange studio. - Required input: exactly two character reference images (starting left and starting right); JPEG, PNG or WebP, each at most 10 MiB. - Built-in input: the site's performance reference video and fixed prompt. Users do not upload a performance video or edit the preset prompt. - Output/defaults: 15 seconds; 480p; aspect ratio 4:3; audio enabled; MP4 delivery. - Output duration, resolution, ratio, model version and output count are not user-selectable in this workspace. One submitted take creates one result. - The MiniMax delivery path normalizes the provider result to the fixed duration/resolution above. Native provider size is not an additional customer option. - Generated results are account-owned; sign in to view and download your completed files. The site does not provide public access to other users' private results. - No separate public image generator, model-training workflow, arbitrary text-to-video editor or H3 Max tool is offered. Marketing guide pages do not establish additional connected models. The site is independent. A model name does not imply official affiliation or guaranteed fidelity, lip sync or successful generation. Review both identities and the complete result.