Field notes from the world of data extraction.
Articles, interviews and analysis on how data is gathered, used and fought over — written by the people closest to it.
Claude Fable 5 is the new best model for writing scrapers
We ran nine models, including the new GPT-5.6, through our Zyte Scraping Code Benchmark. Claude Fable 5 puts Sol and others in the shade - but it’s pricey, and the best extraction code still depends on the best infrastructure.
GPT-5.6, Fable 5, and GLM-5.2 entered a bar “crawl” and got hit by The Rate Limit
When OpenAI shipped GPT-5.6 and Anthropic's Fable 5 was sitting at the top of the price list, my question was not "which one wins a leaderboard." It was the one I actually pay for: for the scraping I do, how much model do I need to buy?
Treat your AI skills like software, starting with evals
Most AI skills are never tested — and it shows. Here's how Zyte evaluates scraping skills like real software, catching failures demos miss.
AI generated these Scrapy projects - why I won't ship them
What happens if you let AI create a Scrapy project from just a simple prompt? Here's what I got and what I had to fix.
Harness Engineering, part 2: harnessing a data extraction agent
Point it at a website, tell it which fields you want, get back clean structured records. That's the agent we're designing in this post — and the interesting part isn't the model, it's the harness decisions that make it actually reliable at scale.
Harness Engineering, part 1: what is an agent harness and why it matters
Same model, same weights, zero retraining — LangChain changed nothing but the scaffolding around a coding model and jumped it from 30th place to the top five on a benchmark. That scaffolding has a name: the harness. And it's the part you actually control.
The harness matters more than the model - Podcast EP07
"The model is the engine — but the harness is everything else." In Episode 7, we dig into why the infrastructure layer around your AI model matters more than the model itself, rank the best models available right now, and ask whether the open-weighted revolution is about to make frontier subscriptions obsolete.
Zyte's first Developer Community Meetup: the recap, slides, and recording
AI agents can now write, run, and self-heal your web scrapers, and in Zyte's first-ever Web Scraping Community Meetup we show you exactly how. Live demos, a Claude Code plugin that turns a prompt into production-ready data, and a fireside chat on where AI is really heading.
How to run any model inside Claude Code
I run GLM 5.2 inside Claude Code with the same tools, the same skills, and the same agent loop. Three environment variables is all it takes.
The best agent skill is the one that says the least
More instruction, worse output. Zyte's head of R&D on why telling your agent exactly what to do can blind it to the obvious answer.
Why I'm adding GLM-5.2 to my agentic coding arsenal
Ayan Pahwa ran GLM-5.2 against Claude Sonnet 4.6 and Opus 4.8 on real scraping work: parity on extraction, 8.5x cheaper on planning, slower and wordier.
How to build your first Scrapy extension
My Scrapy project now plays a triumphant fanfare when a crawl finishes clean and a sad trombone when it doesn't, which is also how I finally learned what Scrapy's extension points are actually for.