render-street-interview
Superseded: the video kit now makes street interviews with goose-studio's street-interview style, which uses the video-h3max part, version 1.0.0 (the take), the transcribe-whisper part, version 1.1.2 and the speech-recut part, version 1.0.2 (the cut), the html-frames part, version 1.2.1 (the overlay), the end-card part, version 1.1.1 and the assemble part, version 1.2.0, in the parts folder of this repository. Its recipes are retired in goose-studio. This renderer stays, unchanged in behaviour, because the write-video-ad-script skill runs its script-context step; its scripts still run.
Human version
render-street-interview
Summary. This is the renderer for the street-interview ad format. Product guessing retains its visible object and reveal. Conversation supports mic-only, prepared-sample and visible-task script/prompt previews. Choose a coherent situation and earned ad connection before writing words. People are generated; conversation media delivery remains unverified.