Migrate video prompts between models (e.g. Kling → Seedance)

**Migrating 3 Kling prompts → Seedance 2.0**

**Clip 1**
- *Original (Kling):* "Camera slowly zooms in on a coffee cup, steam rising, cozy cafe, cinematic, 10 seconds."
- *Rewrite (Seedance 2.0):* "Slow push-in on a ceramic coffee cup, gentle steam rising, warm cafe interior with soft window light, 50mm, shallow DOF, filmic grade. Duration 5s."
- *Note:* Seedance handles 5s beats best — I split the 10s shot into one clean 5s clip; chain a second for the full length.

**Clip 2**
- *Original (Kling):* "Drone shot flying over mountains at sunset, epic, sweeping camera movement, 10s."
- *Rewrite (Seedance 2.0):* "Aerial pull-back revealing a mountain range at golden hour, slow continuous drone glide, anamorphic, warm amber palette, cinematic grain. Duration 5s."
- *Note:* Kling's "sweeping movement" is vague for Seedance — I specified "slow continuous glide" since Seedance reads explicit move verbs better.

**Clip 3**
- *Original (Kling):* "Model turns toward camera and smiles, studio, soft light, dynamic motion blur."
- *Rewrite (Seedance 2.0):* "Model turns toward camera with a natural smile, seamless grey studio backdrop, soft key light camera-left, 85mm, subtle motion, no motion blur. Duration 5s."
- *Note (won't translate cleanly):* Kling's "dynamic motion blur" tends to distort faces in Seedance — I removed it. If you want energy, add it in post rather than at generation.