Hey HN, I built LongCat Video because I was frustrated with AI video generators that only produce 3-5 second clips.

Every tool I tried had the same problem - you'd get a cool 5-second video, then have to generate dozens more and stitch them together. The results were inconsistent and the process was painful.

LongCat Video generates minutes-long videos from text or images with stable motion. No more flickering, no more stitching clips together.

Tech stack: 13.6B dense diffusion transformer model, 720p output at 30fps.

Would love feedback from the HN community. What features would be most useful for your projects?