Сборка explainer-видео с озвучкой и монтажом

★ 8.1 · content

higgsfield-video-explainer is a Claude Code skill that builds a complete non-photoreal narrated explainer video from ordered 10-second blocks, generating one voice take via seed_audio and one clip via gemini_omni per block, then assembling them server-side with explainer_video. The seven-phase pipeline covers style selection, topic research, style-key generation with nano_banana_2, narration scripting, prompt writing, voice synthesis, and final MP4 assembly. It supports live CMS style presets, custom style-reference images, mascot and faceless character modes, 16:9 and 9:16 aspect ratios, and optional burned subtitles at 0.05 credit per voiced block. Best suited for turning a topic or source document into an animated narrated video — not for photoreal films, ads, talking heads, podcasts, or editing finished footage.