We published a full walkthrough of both table scraping approaches using capture_dom with BeautifulSoup and using Gaffa's parse_table action, directly covering when to use each and how to get clean, structured output either way. Read the blog.
We published a full walkthrough on scraping the web in JavaScript using Gaffa's REST API. It covers two core actions: generate_markdown for extracting clean, readable content and capture_dom for pulling raw HTML when you need more control. Also touches on dynamic content handling, geo-routing, async mode, and actions like parse_table and parse_json. Read the blog.
We presented a live session at MLH's Global Hack Week Cloud, walking through web scraping fundamentals, an intro to Gaffa, and two hands-on demos, one scraping a Wikipedia article and feeding it into an OpenAI Q&A session using generate_markdown, and another using parse_json to extract structured data from both web pages and hosted PDFs. The session also covered scraping legality, dynamic content, and how parse_json field descriptions can enforce specific output formats. We've also added both demo notebooks to our Python examples repository:
webpage_to_markdown_and_qa.ipynb — Scraping a Wikipedia page and using it as context for an OpenAI Q&A session.
structured_data_extraction_with_parse_json.ipynb — Extracting structured data from web pages and hosted PDFs using parse_json.
Read the recap.
Last updated
