Hey everyone! I’m Yahia, a founder running brand.dev, we offer a Brand API that lets you pull logos, colors, fonts, descriptions, and more from any domain in seconds but that's not what i'm here to showcase.
We just launched our AI Query API [surprisingly a pain to get right] 🎉
What it does: Think of it as intelligent, reliable web scraping in an endpoint, you send a URL + questions, and it scrapes the website (quickly) and returns structured info like:
All without worrying about JavaScript-heavy pages, bot-blockers, headless browsers, or proxies. You just get clean, reliable data ready for your CRM or onboarding flows.
I'm curious: how are you currently handling customer context or web-derived insights in your AI stacks? I'd love your feedback, ideas, or questions.
Here's a link to the api docs if you're curious: https://docs.brand.dev/api-reference/branddev/query-website-data-using-ai
Interesting, is there a pre-defined list of datapoint_name values, or is it free text?
I guess the big question for me is how consistent and reliable the returned data is. Most AI powered "structure the unstructured" tools like this I've seen have been a little too hit and miss, if there's going to need to be a manual review step to check that the returned data is sane it's just about as quick to grab the data manually anyway.
One way or another, many of my projects ends up needing data from the web...i.e. scraping. There are soooo many options out there. Often I find them too heavy/bloated for my needs. Many don't perform as expected, etc. I end up rolling it on my own. Also not always the best idea. Your idea sounds interesting but I am not totally getting it. Let's say I had a list of urls for say, breweries, and my goal was to get email addresses for an outreach campaign. Can I feed this list to your api with the human readable task "Go to each website in this list and collect the email addresses. Return the email address and url it was found on". Does it do something like that?