Magic Hour integrations and automation

Magic Hour bundles many AI media tools behind one API key: text-to-video, image-to-video, animation, lip sync, face swap, AI headshots, image upscaling, background removal, AI images and more. Automating it makes sense when the same kind of asset is needed repeatedly: a talking-head clip for each product update, consistent headshots for every new hire, upscaled versions of each uploaded product photo.

Every generation is a job rather than an instant answer. A workflow submits a request with its settings and inputs, receives a project id, then checks the video or image project until its status is complete, at which point the response includes download links for the output. Video jobs can take minutes, and the downloads should be copied to your own storage, because they are not meant as permanent hosting. Inputs such as source photos or audio are uploaded first through the API's upload URL step.

Each generation spends Magic Hour credits, and the cost depends on the tool and output; each project reports its credits charged, which for video is an estimate until the job completes and is refunded if it fails. A batch run over a sheet of fifty rows can therefore use a lot of credit, so cap the number of jobs per run and record which rows are done. Generated media also varies from run to run even with the same inputs, so plan for a person to review clips before they go public.

Face swap, lip sync and headshot tools work with people's likeness. Only process photos and voices of people who have agreed, and record that consent next to each row the workflow handles.