Seedance 2.0 "video extend" is a cheat code for animated ads. I made this entire ad with just two prompts. He...

Seedance 2.0 "video extend" is a cheat code for animated ads.

I made this entire ad with just two prompts.

Here’s exactly how 👇 

First, I found an image that represents the company I’m creating the ad for.

In my case, it was an isometric image of a burger place. 

Then I asked ChatGPT:

“Write a ridiculous 30-second script in the form of a multi-shot prompt for the company in the image.”

It gave me this:

"Retro burger diner called “666 Burgers” on a quiet street corner at night, neon sign flickering, cars passing by

[cut] wide shot: normal night, cook flipping burgers, customer eating 

[cut] medium shot: lights flicker, all three look toward the door 

[cut] door slams open, a masked man rushes in fast 

[cut] medium shot: masked man pulls out a gun, points at them

[cut] customer drops food and ducks, cashier freezes, cook steps back 

[cut] cook suddenly throws a hot pan toward him 

[cut] masked man staggers, gun drops 

[cut] all freeze for a second, heavy breathing, tension in silence

[cut] Masked guy slowly gets up, confused 

[cut] Cook turns from scared to cheerful and says in excited manner: Welcome to 666 Burgers! 

[cut] Guy on the ground says: Yeah… best burgers in town! 

[cut] Masked guy: …what? 

[cut] Cook flips a burger and says: Hot. Fresh. Made in under 60 seconds.

[cut] Cashier (pointing to menu): Combo deals start at $6.66. 

[cut] Masked guy: …can I get one? 

[cut] Cashier (smiles): Fries with that? 

[cut] Final shot: all three eating at the counter like nothing happened. Text on screen: 666 BURGERS You don’t leave hungry."

Then I opened Dreamina using an Indonesian IP (via VPN)

Selected the Seedance 2.0 Fast model

Uploaded the image, and used the first half of the prompt (since the max length is 15s). 

This is the first output I've got (basically didn't edit at all, used this in my final video) 

Now comes the most important part - the video extension feature.

Instead of uploading an image or a collage, you UPLOAD THE FIRST 15s video you just generated (the video above) and switch settings to "Omni Reference"

Then add the second half of the prompt. 

Seedance 2.0 takes your uploaded video and generates a continuation that matches the style and characters of the original.

How cool is that? No other AI video model does this! 

Then all you have to do is stitch the two videos together using CapCut or any other editor (and make some additional edits if needed)

If you need extra voice lines at the end (like I did), just generate them with ElevenLabs.

That’s it!

Cartoon Hero 3.0 will have its own dedicated Seedance 2.0 module🔥 

Best workflows, practices, tips & tricks.  

Everything in one place.  

Join the waitlist here 👇

https://docs.google.com/forms/d/e/1FAIpQLScAeNHw2o0mRv4XEV3x2FFQT1x55hGvRtM9SOp7oa6ZYR-RpA/viewform?usp=preview