Create images with consistent characters and art styles in Midjourney/Stable Diffusion

As in the title, I want someone to help me set up a system where I can input training images along with prompts and get outputs in the same style using the same characters. I don't mind which platform is used, I want the best possible output. I have tried basic prompting with Midjourney but have not been impressed. As an example, I attach a 'character' with a consistent art style, created by Midjourney. I would like to be able to prompt something like 'dog running through a field chasing a rabbit' and get the output using the dog and the art style in the images. This could possibly use Dreambooth, but as I say I'm open to all platforms. Thanks in advance, Rob