Show HN: Serif – a zero-dependency, DataFrame for Python
TheAerius Thursday, December 11, 2025
OP here. I built this library out of frustration with messy, day-to-day data: CSVs with duplicated columns, APIs returning nested JSON, Excel sheets, and various ad-hoc ETL requests.
Tools like Polars and DuckDB are great for heavy analytical workloads, but sometimes I just want a lightweight, Pythonic table I can iterate over. I want the ability to for `row in table:` without surprising semantics.
Serif takes a vector-first, zero-dependency approach aimed at everyday data tasks.
https://github.com/CIG-GitHub/serif
This is an early release, so I'm looking for feedback on ergonomics and places where the API is either missing features or could be simpler.
Summary
Serif is an open-source tool for extracting structured information from unstructured text. It utilizes a combination of machine learning and rule-based techniques to identify entities, relationships, and events within documents, enabling users to generate insights from complex data.
3
0
Summary
github.com