BrowserBook
The Browser Automation IDE
Start new thread
Chris Schlaepfer

Adding configurable proxies for API-based execution

by
BrowserBook
•

Hey PH - today we’re rolling out configurable proxies for API-based execution in BrowserBook.

One of the most common reasons browser automations fail in production is bot detection. Captchas, IP reputation issues, geo restrictions, and rate limits can derail even solid Playwright workflows once you start running them at scale.

This update gives you explicit control over how your automations run in production.

What’s new:

Configurable proxy support: you now have three proxy type options depending on your needs:

  • Datacenter proxies: Traffic routed through commercial datacenters. Fastest and most cost-effective option, best for sites without aggressive bot detection.

  • ISP proxies: Traffic routed through datacenters using residential IPs leased from ISPs. A middle ground between speed and detectability, with more trust than pure datacenter IPs.

  • Residential proxies: Traffic routed through real residential IP addresses. Least detectable option, useful for sites with stricter bot detection or geo-sensitive flows.

The goal is simple: make API-based execution behave more like a local browser session, without sacrificing speed or reliability.

Proxies are entirely opt-in, and all remote executions still use stealth-mode functionality, including Kernel's built-in CAPTCHA solving.

45 views

Add a comment

Replies

Best
Bhavin Sheth

This is a very real pain point — most automations don’t fail because of logic, they fail because production environments don’t behave like local ones.

Giving explicit control over proxy type feels like the right abstraction instead of hiding “stealth” behind magic defaults. Datacenter → ISP → residential as a conscious tradeoff is especially nice for teams scaling gradually.

Curious to see how teams decide when to switch tiers in practice. Solid update

Thomas Jack

This is honestly super helpful. I've had solid Playwright flows break the moment they hit production, mostly due to IP issues. Having explicit proxy options lets me tune execution instead of guessing. The opt-in approach is nice too, flexibility without forcing extra complexity.