Patrick van den Bosch
Founder, Hyper Solutions
Patrick founded Hyper Solutions and builds its anti-bot APIs. He reverse engineers bot protection systems such as Akamai, DataDome, Incapsula and Kasada, and writes about request-based scraping: what these systems check, how they change, and how to keep a scraper working without a browser.
Articles by Patrick van den Bosch
Scraping API vs Web Unlocker vs Anti-Bot API: Which Architecture Fits Your Scraper?
Three ways to get past anti-bot protection: a scraping API, a web unlocker or an anti-bot API. The real difference is who owns the HTTP client, and that decides control, debugging, proxies, cost structure and lock-in.
Why Rotating Proxies and User Agents Can Get Your Scraper Blocked Faster
Rotating the IP and User-Agent on every request is standard advice for scraping without getting blocked. On modern anti-bots it often backfires, because they judge a session, and a session that changes identity on every request contradicts itself. Rotate between sessions, not inside them.
Python Requests 403 Forbidden: How to Find Out What Is Blocking You
Python requests returns 403 Forbidden on a URL that works in your browser or in curl. Read the response first: tell an application 403 from an IP block, a client fingerprint problem or an Akamai, DataDome, Incapsula or Kasada challenge, then take the right next step.
Akamai SBSD and DataDome: Three Fields Become Required on 8 December
From 8 December 2026, every Akamai SBSD call must send scriptUrl, every SBSD call after the first must send context, and DataDome interstitial and slider requests must send the challenge script. Errors ramp up from 1 December. Here is what to change and how to check you are ready.
Kasada iOS SDK i-2.x: x-kpsdk-dv Now Carries the App ID and Version
On iOS, Kasada sends an x-kpsdk-dv header that describes the device. Since SDK i-2.x it also names the app, and a header without the app fields is rejected. What the header contains, how to build it, and how to pass it to our API so the payload matches the device.
New SDK Majors: Akamai Telemetry, Client Hints, SBSD scriptUrl and Kasada strict
Go v4, Python 4.0 and JS 5.0 are out. Every generate call now returns a result object, the APIs return the full set of Accept-CH client hints, Akamai gets a telemetry endpoint for sites that protect their API with an Akamai-BM-Telemetry header, SBSD takes a scriptUrl and Kasada takes strict.
Stop Uploading Challenge Scripts: scriptHash, and Kasada x-kpsdk-v from p.js
Upgrade the SDK and your DataDome, Incapsula and Akamai requests send a 64 character hash instead of a 600 to 800 KB challenge script. No code changes. Plus a Kasada endpoint that reads x-kpsdk-v from p.js.
Kasada Mobile SDK v2: The User-Agent Is Now an MD5 Hash
Kasada's v2 mobile SDK no longer sends a readable user-agent. It sends a deterministic MD5 hash of the OS version, app bundle ID and app version. Here is what changed, what goes into the hash, and how we confirmed it against live traffic.
How to Bypass Incapsula (Imperva) Without a Headless Browser: Request-Based Guide
Bypass Incapsula (Imperva) bot protection over plain HTTP: detect reese84, dynamic reese84 and ___utmvc, renew tokens on time, and fix captcha blocks.
How to Bypass Kasada Without a Headless Browser: Request-Based Guide
Bypass Kasada over plain HTTP: spot the 429 + ips.js challenge, post to /tl, handle x-kpsdk-ct, st and cd tokens, and debug 429s without a browser.
How to Bypass DataDome Without a Headless Browser: Request-Based Guide
How a request-based DataDome bypass works: identify the challenge, keep TLS, headers, cookies and sessions consistent, and debug blocks without a headless browser.
How to Bypass Akamai Without a Headless Browser: Request-Based Guide
How a request-based Akamai bypass works: keep TLS, headers, cookies and sessions consistent, and debug blocked requests without a headless browser.
Akamai SBSD Adds a Context Field: What to Change
The Akamai SBSD endpoint now returns a context field alongside the payload and accepts it as an input. Keep the context from your first SBSD response and send it on every call after that, with the script field left empty.
DataDome Is Moving the Challenge Script Out of the Page: What to Change
DataDome has started serving the challenge script from its own file instead of inlining it in the challenge page. Here is how to detect it, what to fetch, and what to send so your payloads keep matching the script the page actually served.
How to Use powhttp: Recording, Comparing, and Proxying Traffic for Request-Based Scraping
powhttp barely has docs, so here's the guide. What it is, why request-based scraping needs wire-level truth DevTools can't show, and how to record scripts and browsers, search a session, diff requests 1:1, replay them, route through external proxies, and export a HAR our analyzer reads.
One Prompt, a Full Airline Scraper: Announcing Our Claude Code Plugin
We let Claude Code one shot a complete request based scraper for Etihad's flight pricing API in 16 minutes, then ran it with zero AI in the loop. Here's the demo, the numbers, and our new Claude Code plugin.
Header Order Is Not Enough: Why Bad DataDome Captures Still Get Blocked
Most request-based DataDome failures come from copying headers out of a browser session that was not recorded cleanly.