As you probably know, I've been generating wam video content using Kling AI which still amazes me. But I have one frustration which annoys the hell out of me.

Suppose you upload a 'racy' image to convert to video. Kling won't accept it. Fine, I can live with that. But suppose Kling accepts it, and your prompt goes through. The 5 second video is created, then you click to add 5 more seconds to it, and it is lengthened. All good so far. You add 5 more, then 5 more again will you're up to 20 or more seconds, even 30 or more.

But the next attempt to add 5 more seconds results in 'Process Failed - Change Prompt'. This means you've paid (with your credits) for all those time extensions and you've seen the results, and you're excited to find a spot where you can stop adding and save this 30 or 40 second scene. And they had been ok with the prompt all along, using your credits too. But this error message means you must start over from the beginning and they have tossed out all of the paid-for time extensions you've watched being created.

Even if you started over and spent more credits, there's no guarantee that the results would be anything like what you had the first time.

Has anyone else experienced this? I don't yet see much video content being created so maybe not many are even using Kling yet.

I actually have not experienced this. However, I haven't made anything longer than 20 seconds. That said, I created and downloaded my initial 10 second creation. Then I added 5 seconds and downloaded that, then did the same with the next 5 seconds. It also treated each addition as a whole new generation and I had each stage as a seperate video in my history.

I have now begun to download each incomplete video until I have the last one done. What gets me is they accepted the initial prompt, yet say they have a problem with the prompt 3 or 4 versions later. Thanks for your input!

You just have to watch what you're doing fits their terms of service I guess. They are looser than Microsoft but it's not supposed to be NSFW.

You can also download the video after each extension, so you have copies of 5s, 10s, 15s, etc saved locally.

You could try manually grabbing the end frame of a video (using say VLC or your video editing software) and start a whole new generation with a new prompt - and then you can stitch this together manually.