⬢github Python · 10 ★ +4 since we first saw it · pushed 21 min ago · MIT
nicobrenner/jeffy
Pretrained text classifiers you can run and retrain on CPU
Jeffy is a Python package that ships 13 pretrained text classifiers (spam detection, sentiment, intent, news topic, etc.) that run entirely on CPU using logistic regression over a shared embedding encoder. It includes a local web UI and REST API for predictions, plus a CLI to train and serve custom classifiers from your own CSV, TSV, or JSONL data — no GPU required.
Why now: It was recently launched on Hacker News as a 'Show HN' demo, pitching the angle that local pretrained classifiers don't need a GPU.
Who it is for: Developers who want lightweight, self-hosted text classification (spam, sentiment, intent routing) without GPUs or cloud APIs.
Stars over our 7 snapshots: 6 to 10, since 1 h ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/nicobrenner/jeffy