The idea is to have deterministic text transformations generated by AI, which can be saved and reused later to avoid incurring/piling up LLM-call cost and hallucinations.
This solves my own pain points when I wanted to clean up some text with non-trivial requirements which takes quite some time to figure out with normal text editor's find-and-replace + regex.
Yes, you can write your own code/put LLM generated code in an IDE and do the same thing. However, I just wanted to create a trivial way to generate/save/execute (safely in a local lightweight VM) code to transform my text.
The idea is to have deterministic text transformations generated by AI, which can be saved and reused later to avoid incurring/piling up LLM-call cost and hallucinations.
This solves my own pain points when I wanted to clean up some text with non-trivial requirements which takes quite some time to figure out with normal text editor's find-and-replace + regex. Yes, you can write your own code/put LLM generated code in an IDE and do the same thing. However, I just wanted to create a trivial way to generate/save/execute (safely in a local lightweight VM) code to transform my text.
Would love to hear some comments and feedbacks!