fenic Unveils Semantic DataFrame Combining LLM Operations and SQL-Style Queries

fenic is a DataFrame query engine that integrates PySpark/SQL-style operations and language model (LLM) calls within a single query model to process unstructured data such as documents, transcripts, and logs.

Unlike existing data processing approaches, it allows developers to mix standard operators like select, filter, join, group_by, and agg with semantic operators powered by LLMs. This enables developers to perform natural-language-based data extraction and transformation using SQL-like syntax, without building complex pipelines.