Seedance 2.0 is Jmeng’s multimodal creation experience: image, video, audio, and text in one go. This seedance manual shows four simple seedance guide workflows so beginners can get results fast.

Seedance 2.0 beginner

Multimodal and all-in-one reference

You can mix up to ~12 files; prefer the assets that matter most for the shot and rhythm. In Jmeng choose Seedance 2.0 + All-in-one reference, up to 15s. Always use @ in the prompt to point to assets—it’s required.

Play 1: Fight scene

Upload character and scene images; in Jmeng: Video generation → Seedance 2.0 → All-in-one reference. Example prompt:

“@image1 and @image3 characters fight in @image2’s scene; intense, cool action; @image3’s character defeats @image1’s in the end.”

Play 2: Character replacement

Use a ~15s video and one character image. Prompt example:

“Replace the character in @video1 with @image2’s look. Keep @video1’s camera, transitions, composition, mood. Match motion to the lens, one continuous take.”

Use a clean background on the character image for better blend.

Play 3: 4-panel anime

Use a 4-panel manga image and one anime clip as reference. Example:

“Animate @image1 top to bottom; keep dialogue consistent with the panels. Add cuts and sound for key moments; gentle, detailed style. Reference @video1 for the style.”

You can extend to 9-panel or more.

Play 4: Scripted video

Write a script (e.g. with Doubao, Kimi, or DeepSeek) in table form and screenshot it. Upload to Jmeng, Seedance 2.0, All-in-one, matching duration. Example:

“Follow @image1’s storyboard. Match its shots, framing, camera, visuals, and sound; produce a video about [topic].”

You get a video that follows the script.

These four plays cover the most common seedance manual cases. With @ and all-in-one reference, you can follow this seedance guide and experiment. Use the button below to start.

👉 Start with Seedance