A · Local animation
We animate Globi with reusable character layers and draw the props locally. A saved lesson plan controls gestures, captions, and learner pauses. We use ElevenLabs’ Cristina - Premium Spanish Female (voice ID 1CeqBeXMOqCleeQjfYfO). The same assets and recording let us reproduce the lesson on the Mac.
B · Google Flow with built-in voice
We give Google Flow a Globi reference image and a lesson prompt. Its Omni Flash model generates the animation and Spanish speech in one pass. We can reuse the setup, though each new take can look or sound different.
B · Google Flow with ElevenLabs voice
We reuse the Google Flow footage, adjust its timing locally, and add the same ElevenLabs narration used in A and C. That lets us compare the generated movement with the same voice and learner pauses. We can revise the local edit without spending more Flow credits.
C · Local animation with AI-generated artwork
We keep the same local Globi rig and lesson timing as A, and use AI to create supporting artwork such as the map or cheese. We save each illustration so future renders can reuse it on the Mac.