EdgeSpeak Video Voice

Make a video that Claude builds in code speak, and make every caption word land exactly when it is said. The whole chain runs on-device: the narration text and the audio never leave the machine.

narration script ──speech──▶ narration.wav ──align──▶ words.json ──video_cues.py──▶ cues.json / timing.js
                                                                                        │
                                            HyperFrames · Remotion · Canvas · Manim ◀───┘

This skill composes two others: edgespeak-broadcast (speech synthesis) and edgespeak-align (word timestamps). It links to them for the full model / voice / error tables instead of repeating them; read the relevant section there whenever a step below fails.