faceless-explainer

Provide your article text to the faceless-explainer workflow, which plans a storyboard, writes a narration script, generates voiceover and music, builds each frame as an HTML composition, and renders a final MP4. No footage or website capture is needed since all visuals are invented per scene.

A faceless explainer is a video where every visual is invented downstream — typography, abstract graphics, diagrams, and data-viz — rather than captured from a website or real footage. It suits topic explainers, concept breakdowns, how-tos, and listicles.

Yes. Add brand colors and fonts to capture/extracted/tokens.json before Step 2, and the build-frame script remixes a design preset onto them by role, clamping type weights to the faces your font actually ships. Without brand tokens, the preset's complete palette is kept.

Yes. Set music: none in the storyboard frontmatter and omit SCRIPT.md to mark the project fully silent — no narration, BGM, or SFX is generated. Using music: none with a script keeps text-to-speech but disables only the background music.

Small 1–4px text_box_overflow findings on caption highlight words are a known false positive: the caption pill uses a snug line-height with no overflow clipping, so glyph ink spills into its own padding without visible clipping. Only act on overflow findings naming frame elements, not caption words.

Use product-launch-video when the video is a promo or tour built from an actual website that can be captured. Faceless-explainer is for explaining a topic from text alone, where there is no site or footage to capture and every visual must be invented.