The faceless-explainer skill is a specialized workflow for creating synthetic explainer videos. Unlike captured-asset workflows, it relies on synthetic assets (typography, abstract graphics, diagrams, data-viz) rather than real-world footage.
Supported Formats
Formats are derived from the destination and must be set once in the STORYBOARD.md frontmatter:
- Landscape:
1920x1080 - Portrait:
1080x1920 - Square:
1080x1080
Workflow Differences
- No Capture Step: Uses synthetic
tokens.json and visible-text.txt instead of Step 1 capture. - No Asset Staging: Does not use
asset-descriptions.md or capture/assets/. The asset_candidates field is empty by default. - Visual Generation: Every visual is invented by Step 5 workers (typography, abstract graphics, diagrams, data-viz).
- User Assets: The only real asset path allowed is a user-supplied image in
public/<basename>.
Available Background Scripts
These scripts are located under scripts/ and supplement the hyperframes CLI:
build-frame: Adopts and brand-remixes a frame preset into frame.md (includes caption skin).audio: Handles TTS, transcription, BGM, SFX, and duration syncing.captions: Manages captioning.transitions: Injects and verifies transitions.assemble-index: Final assembly step.