AI video models are becoming more sophisticated all the time. Veo 3 can already produce highly realistic videos. In MeiGen-MultiTalk, we have an open source framework for creating synced video content where multiple characters talk and sing. As you can see on the project’s page, this model can generate high realistic singing videos. It works on cartoon videos too.
To pull this off, you are going to need an audio file, reference image, and a prompt. This model has 480p and 720p output. It supports videos up to 15 seconds. It supports single and multi-person generation.
