On Wednesday, we met up with @taylordowns2000 and @jclark to do a one-day OpenFn AI hackathon.
We focused on improving the user experience when reading the OpenFn documentation. To assist users who might need additional context to understand some parts of the documentation or tutorials, we built an LLM-based tool that can generate an explanation for any paragraph on the documentation site.
The tool gathers relevant context from the documentation through Retrieval Augmented Generation (RAG) and queries a Claude model to clarify the paragraph.
The tool isnβt launched just yet β we need to do some more robust testing first. For now, you can watch a demo on Loom or check out the code here.
This project builds on @satyammattooβs excellent work on the RAG Search service. Thanks Satyam!