Omni Video2Note

Use omni_video2note_create with an absolute local video_path and a .pdf output_path.

The tool runs a bounded workflow internally:

1. Understand actions, speech, visible text, and timestamps with Omni. Long videos may require several chunks.
2. Use the same Omni model once more to produce the note and step timestamps together.
3. Extract screenshots near those timestamps locally and render a searchable PDF. There is no separate model selection, format-repair, or final PDF-review request.

Guidelines: