From the HackerLinks archive

Pulpie

Pulpie is a model-driven web cleaner that preserves main content, tables, and images.

At a glance:

First seen:2026-07-06

Last seen:2026-07-06

Times seen:1

Website:huggingface.co

The short version

Pulpie is a model-driven web cleaner that preserves main content, tables, and images.

Why it caught our attention

The thread confirmed it does useful extraction work and compares favorably with heuristic cleaners.

Where it surfaced on Hacker News

2026-07-06

Editorial paraphrase

The author said images and tables pass through, and that the model outperformed heuristics by 25 F1 points on their tests.

Show HN: Pulpie – Models for Cleaning the Web