Scraping to feed a RAG
A step-by-step guide to turning public documentation into clean text, chunking it by section, and putting it into a semantic search without writing a custom scraper.
Blog
Reviewing 115 services one by one turns up things the pricing page never says. They are written down here, so they save you an afternoon.
16 articles Page 1 of 2
A step-by-step guide to turning public documentation into clean text, chunking it by section, and putting it into a semantic search without writing a custom scraper.
The four questions that decide whether a scrape is defensible: public data, terms of service, personal data and copyright. Plus what never to do.
What web scraping is, when it beats an official API, and the five decisions that determine whether your scraper lasts a month or breaks on Tuesday.
The six pieces a scraper needs — HTTP client, parser, browser, queue, storage and alerts — and which options have a genuinely free plan for each.
ChatGPT, Perplexity and Google's AI summaries cite specific sources. What makes them pick yours: direct answers, HTML without JavaScript, structured data and an llms.txt.
The seven pieces you need to measure and fix the SEO of a small project without paying anything, and what to check in each one's free plan before you set it up.
How to get Google to index both versions of your site instead of treating one as a copy of the other. With the canonical mistake that erases an entire language from the index.
JSON-LD without the hype: which types actually do something, how to stop the schema promising things the page doesn't show, and why it doesn't raise rankings but does raise clicks.
A step-by-step guide to a sitemap that lists exactly what is public and a robots.txt that blocks only what is private. With the three failures that break nothing and drop you from the index.
115 services with a genuinely free plan, each one with its limits spelled out. No sign-up needed to start looking.