Image to Video Generator (v3.0.0 – Working Veo-3.0 REST API)

Status: ✅ Tested & Working
Last Updated: 2026-04-11
Example: Golden Tibetan offering → 2.4 MB video in ~60 seconds

What it does

Generates cinematic 5-second video from any image using Google's Veo-3.0:

1. Gemini Vision (2.5-flash) analyzes image for motion/scene description
2. Veo-3.0 predictLongRunning generates video asynchronously via REST API
3. Polling loop monitors operation until video is ready (~60-90 seconds)
4. Download saves MP4 to ~/.openclaw/workspace/tibetanProc/ with yymmddHHMM prefix

Key Fix (v3.0.0): REST payload must use bytesBase64Encoded field (not data or inline_data)