This is a super simple script to parse output from stable diffusion (automatic1111) and generate a vault with interconnected nodes, corresponding to the words you've used in your prompts as well as the the dates you've generated on.
- Add a way of visualizing strength of prompt word association
- Utilize actual clip parsing to work with tokens rather than raw words
- Something more advanced to relate things in latent space? (I'm not smart enough to know if this is a real thing)