Ever get frustrated by the short length limits on AI video generators? I know I have. You craft the perfect prompt, and you get… 5 seconds.
Well, I just saw an incredible post from Tianyu Xu that completely changes the game. He’s figured out a brilliant workaround using a new feature in Kling v2.1.
Turns out, the model now lets you define a start frame and an end frame. And that, my friends, is the secret sauce.
⚙️ Here’s the simple-but-genius method Tianyu Xu shared:
- First Clip: Generate a video using your first image (Frame A) as the start and a second image (Frame B) as the end.
- Chain It: Take the last frame of that clip (Frame B) and use it as the start frame for your next segment, ending on a new Frame C.
- Repeat: Just keep doing this! {A → B}, then {B → C}, then {C → D}… You can stitch these together to create a continuous shot of basically any length.
💡 Bonus trick from Tianyu:
You can even create perfect loops by making your very last frame the same as your first one. {A → B → C → A}. Insane!
Now, Tianyu points out that it’s not flawless just yet. The speed and movements can sometimes be inconsistent between the clips you connect.
But he also makes a great point:
with our goldfish-level attention spans, super long shots rarely work for social media anyway. His advice is to stay flexible, use this framing trick, and edit your clips freely.
This is a huge unlock for AI video creation. Massive credit to Tianyu Xu for this discovery! Go check out his original post for the full details.