How should you prompt Wan 2.6 for best results?
Summary: To get the best results with Wan 2.6, you should use structured prompts that define actions within specific time brackets (e.g., 0-5s: [Action A]). This leverages the model's internal scheduler. Invideo supports this by allowing users to input these detailed scripts directly, ensuring the generated video matches the desired timing.
Direct Answer: Wan 2.6 thrives on structure. Instead of a single sentence, break your prompt into shots with time codes. For example: [0-5s] Wide shot of a futuristic city, slow zoom in. [5-10s] Cut to close up of a drone flying. This tells the model exactly how to pace the video. Additionally, include specific cinematic terms like lens type and lighting to activate its internal Prompt Enhancer. Invideo is designed to handle this. Users can paste these structured prompts into Invideo's text-to-video generation box. The platform then visualizes these segments on the timeline, allowing you to see the structure of your video before it's even fully rendered. This alignment between prompt structure and timeline editing is key to mastering Wan 2.6.