The web seems to be becoming paradoxically less and less machine/automation friendly. I recently had to modify 150+ accounts in a Google Workspace. Twenty years ago, I would be doing this in a Unix environment with a very simple shell script. Instead, I had to click-copy-paste-click, 20 times per page, like a monkey. I'm sure there must be some sort of API, but it would have taken 100 times the time it would have tak…
Less distributed/democratized machine/automation friendly... But definitely automation friendly for giant search scrapers, with lots of compute, cash, IP blocks and AI.
I'd like to change that. I originally created BrowserBox^0 as a platform to serve "web scraping authoring tools". These tools are normally served as extensions, or even downloadable electron apps. But what about something easier to distribute, more powerful, more lightweight, and less beholden to walled-garden gatekeeping? BrowserBox changes all that, as it's clientless and runs in a regular web browser even on mobile. Anyone can build a scraping script on top of it, even from your mobile device while riding the bus. That's the vision anyway. But I got side-tracked by how the "embeddable browser" is a useful product in its own right. I still intend to return to fulfilling its original purpose however.
The key is to build a good "extensions-like"-but better-API atop the Chrome DevTools protocol and our BrowserBox functionality. We're open source so come visit if you'd liked to get involved or check it out! :)