Pinned Tweet
Learn how to bypass Cloudflare Bot Management. You'll add evasions to skip blocks by understanding how it works and what sensor data it sends: zenrows.com/blog/bypass-clou… #webscraping #antibot
3
4
20
Two days at #AISB26 with 10,000+ builders. Our biggest takeaway: agents are going into production, and reliable web data is where they break first. That is the problem we solve. Thanks, Barcelona 🧡
1
1
5
93
One command adds the Zenrows MCP server to Claude Code:
claude mcp add zenrows -e ZENROWS_API_KEY=YOUR_ZENROWS_API_KEY -- npx -y @zenrows/mcp
1
1
80
Adding Zenrows to your agent is simpler than you think.
Hit "Add to your agent" on our homepage and paste the prompt into your agent. It asks how you want to connect, sends you to a free key, writes the config, and runs a test fetch.
Try it on our homepage 😇
1
2
69
Who's going to Barcelona? Catch us at BCN!
aisummitbarcelona.com/
3
63
How do you know when a scraping job needs a browser? 🤔
Three sources for our World Cup dataset. Two worked with a plain Fetch call.
A page built with JavaScript doesn't always need a browser. Session tokens are what actually force you into one.
zenrows.com/blog/scrape-2026…
1
9
94
Which combo would you need to build a RAG pipeline over 50 doc sites? 🤔
Half are JS-rendered. A few sit behind a login. Content changes weekly.
Four combos on the card. Pick one, or tell us what we left out.
1
2
126
1. Half these sites will return an empty shell instead of the content.
zenrows.com/blog/ai-web-scra…
2. Some of these will return a 200 and the wrong data.
zenrows.com/blog/best-web-sc…
3. Most of these work fine until the content changes.
dev.to/zenrows/how-to-build-…
35
Reintroducing you to the Zenrows Agent Toolkit!
It's open source, MIT-licensed, and one command wires Zenrows into your IDE, MCP clients, coding agents, and SDKs.
Free plan comes with 5,000 credits a month.
ordnl.link/HWZu7qk
1
1
3
127
Here's how to collect a fine-tuning seed from sources that block standard crawlers.
Set mode=auto and response_type=markdown for protected pages.
Switch to response_type=pdf with js_render=true for filings and papers.
ordnl.link/wFxnXvn
1
2
76
We just scraped the 2026 FIFA World Cup data... 'cause we can 🙄
- FIFA's Golden Glove winner isn't in our top 10 keepers (awkward)
- Japan had the best conversion rate shooting from outside the box
- Argentina led in bookings, got zero VAR calls (no comment)
All 15 findings:
1
1
3
104
Start using Zenrows Fetch as a typed tool inside AG2 🔥
AG2 agents share one history. Without a fetch tool, the researcher comes back empty and every agent after forwards that.
Register Zenrows and clean Markdown lands on the first call.
ordnl.link/kH4vfsP
2
85
Zenrows Fetch can run as a function_tool inside the @OpenAI Agents SDK.
Register it alongside the built-in WebSearchTool and your agent gets two ways to reach the web.
Search finds the URL;
Zenrows reads the live page.
Setup and code: ordnl.link/gRg807y
1
77
This was our stand at DevBcn 2026.
The banner said it better than any product page could. Developers stopped, read it, and started conversations we didn't have to open.
That's the version of Zenrows worth knowing.
If you missed us in Barca, the infrastructure's still here.
1
3
102
Hidden tip for this week.
Six sellers listing the same product means six different title formats, currencies, and variant structures to reconcile.
Zenrows Marketplace Catalog fetches, normalizes, and dedupes everything in one run.
Run it free: ordnl.link/5T0tsYs
3
84
We tested the open-source Puppeteer MCP for AI web scraping.
It worked on unprotected JavaScript-rendered pages but failed on protected sites.
Full breakdown, test results, and setup guide here zenrows.com/blog/puppeteer-m…
4
249
If you have ever had a scraper report a 100% success rate and collect nothing, this is probably why: a lot of block pages come back as HTTP 200.
The request succeeded. The response is a challenge page. Your pipeline stores it and moves on.
1
2
171
Checking the status code is not enough. Assert on the content you expected to find, and a soft block turns into a loud failure instead of a silent one.
1
35
That's one of 70 questions we answered at zenrows.com/glossary, covering how pages get fetched, why sites refuse them, the errors you hit, and how AI agents read the live web.
8
GianfrancoGugino once said this.
The job changed underneath the word. An AI agent hitting a page mid-task can't stop to debug a broken selector. It needs clean data right now to finish the job.
Which tool would you reach for this week?
2
54
If you're trying out Browser Sessions, read this 👇️
Browser Sessions excels at handling complex scenarios that traditional HTTP-based scraping can't address.
1
1
102
- Run custom JavaScript inside the page context to extract computed values, trigger events, and process data before it leaves the browser.
- Generate PDFs from dynamic pages with custom headers, margins, and print backgrounds without a print dialog or local browser.
1
20
See use cases and example code here: docs.zenrows.com/browser-ses…
11