Hacker News Reader

I ended up building an MCP that pulls results from multiple places and dedups it.

The free tiers on most are generous enough for normal agent MCP usage

Also...pretty sure decodo just launched a search thingie

One thing I see on the new gen of ai search bots are they fake chrome and rarely disclose themselves in UA, and not respect robots.txt, and no rate control for smaller sites.

I’m not accusing any one engine discussed in this thread, but hearing a proxy vender like decodo providing search give me a feeling there’s non zero over lap to bad search and residential proxy crawler.

Think the era of respecting stuff like robots.txt is over.

I build a web-fetch tool for my own LLMs that is essentially the same tech as web-scraping except its single requests not high volume.

You can't really get anywhere these days without disguising UA etc.

Real human hit once with disguised ua is fine, but I’ve seem a few million clearly batch crawl request within a few minutes hitting my potato server from thousands of IP addresses, and they do it every four weeks almost at same hours. cloudflare doesn’t help much if at all, I can do proof of work but it creates problems for real human, I can rewrite in rust with some sophisticated cache but i don’t want to.

Anyway I’m not expecting a solution. Just say this sux.

Reply on news.ycombinator.com