Hey HN! We built a user-friendly tool that allows non-technical domain experts to explore and evaluate the effects of LLM-generated prompt chains on large datasets (via CSV). Our solution wraps around the ChatCompletions API (gpt-3.5-turbo), offering an accessible interface for users who lack the skills to work with Jupyter notebooks or other tooling.

The current feature set is minimal, tailored to our friend's specific needs. We're eager to improve and expand the tool, so please share your feedback and suggestions--brutal honesty is okay!

Note--we have the space for the OpenAI API key because we don't know how many people will use this and don't want to run up our GPT bill . Sorry for the inconvenience for those who don't have a key :((

Demo video: https://drive.google.com/file/d/15VAjHWt7Btgutp5pc6QPnL8YfLO... Initial CSV from demo: https://drive.google.com/file/d/19LUrXBZ08mIo7p0xJKTwOMdEle8... Result CSV from demo: https://drive.google.com/file/d/1f-kzSvhgcg6Bc0wCODomPl1FwCM...