brand-video

name

brand-video

description

Produce one or many consistent, on-brand faceless marketing videos end-to-end from a topic — script → ElevenLabs voiceover (with timestamps) → one styled illustration per beat via the grounded pipeline (generateImage(), auto-detect + auto-LoRA) → ffmpeg stitch to 1080p, optionally uploaded to YouTube as Unlisted. The visual LOOK is a selectable style from styles.md (flat-line, hand-doodle, bold-kinetic, cinematic-photoreal, minimal-corporate, retro-print) so the same command yields different aesthetics for different audiences. Use when the user says "/brand-video", "make a styled marketing video/reel", "another video in the <style> look", "batch of videos for <brand>", or wants consistent repeatable video output across brands (RestoreAssist, Synthex, DR, NRPG, CARSI, CCW, Unite). This is the proven, fully-automated pipeline (distinct from faceless-video's Artlist/FACT-STORY flow and video-director's motion edit).

brand-video — consistent styled faceless videos, on demand

Invoke as /brand-video. One command, many looks. Built from the validated RestoreAssist run (9 videos shipped). Picks a style from styles.md, holds everything else constant so output is consistent batch-to-batch and brand-to-brand.

Inputs (ask only if missing; infer sensibly)

• brand — RestoreAssist / Synthex / DR / NRPG / CARSI / CCW / Unite, or a client. Sets claims + CTA + voice.
• style — a key from styles.md (default flat-line). This is the "look" / dropdown value.
• topic(s) — one angle, or a list for a batch (one video each, parallel).
• voice — ELEVENLABS_VOICE_ID (per-brand; from ~/.hermes/.env). CRITICAL: confirm the current value before generating — the user changes it.
• upload? — default: render + stage only. Upload to YouTube Unlisted only when asked.