Show HN: fenic – LLMs as dataframe operators, query meaning and structure

Category: library

Tags: dataframe, llm, semantic-processing, pipeline, python

Score: 7.5/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 7)

fenic is a semantic DataFrame engine that integrates LLM operators like extract, classify, and summarize directly into query pipelines, enabling reproducible and inspectable processing of unstructured data. It combines relational operations with AI-powered transformations, allowing pipelines to be reused by both humans and agents via MCP tools. The project stands out for making AI-assisted data exploration a durable, typed artifact rather than a one-off script.

Target audience: data engineers, data scientists, AI engineers, backend devs

Repository: https://github.com/typedef-ai/fenic · Python · Apache-2.0 · 461 stars

View on Hacker News