Most AI video tools show you a gallery on the marketing page and something different when you pay. The only honest way to sell a single-effect tool is to publish the output itself, which is why every clip here is a file the live generator returned rather than a mock-up or purchased footage.
Look at the three of them next to each other and you will notice they share a scene, a camera position and a clapping rhythm. That is the fixed template doing its job: the photos supply who is in the clip, the template supplies everything else. It is also why two CatClap clips are recognisable as the same meme rather than as two unrelated videos.
The three examples were picked to cover the cases people ask about before buying: a black cat (the hardest source photo), a patterned cat (the easiest to check identity against) and a dog (proof the workflow is not cat-only). Read the note under each one and you will know what to look for in your own first render.