Prompt Extraction

---
title: "Prompt extraction (GUI/script)"
description: "Learn how to use a large language model (LLM) to build a simple GUI for prompt extraction."
social:
  cards:
    image: https://docs.bydanielrosehill.com/repos/prompt-library/images/sloth-and-sieve.webp
og:
  title: "Use An LLM To Create A Prompt Extraction GUI!"
  description: "Learn how to use a large language model (LLM) to build a simple GUI for prompt extraction."
---

# Use an LLM to code a prompt extraction GUI!

Rather than pretend that I know more about Python programming than I do, I'll be honest and say that I'm a Python imposter. 

While I strive to improve my understanding of the syntax, many scripts and GUIs that I slap together are mostly generated by .... you guessed it ... LLMs.

The good news is that many of the scripts that LLMs author actually .... work (I'm still amazed by this).

Exhibit A: a Python GUI for extracting prompts from an Obsidian vault containing prompts and outputs jumbled together in loosely delineated templates:

While it's a good idea to keep a separate repository of your prompts and to curate it intentionally, sometimes you may wish to extract prompts from a big hunk of markdown files that contain prompts and outputs hopelessly jumbled together in an admixture of data. 

But even if you've denoted `Prompts` and `Outputs` you can usually prise them apart.

Try something like:

## A prompt to create a GUI for extracting prompts

Generate a Python GUI.

The program will go through a directory of markdown files.

Its base directory is:

`{path}`

The program should recurse through that base directory and process every markdown file (.md, .mdx).

The program will find markdown files that follow this template: