for now, my benchmark is a simple wrk test with varying connection counts (100, 256, 1000). Web server components like the parser, i/o, and the server itself have their own independent benchmarks (rust built-in), along with some e2e tests (mostly written in Bash).
We might be interested in the technology (not as a VC), if it shows meaningful improvement (>2x) over what we currently do.
If you could show how it compares with a simple replicable baseline, maybe something like C program that just accepts a connection? Lots of these things are hardware-dependent.
Right now we use monoio and have a draft benchmark with speed. Happy to continue talking over e-mail. Should I write to you?
Hey! I'm Evgeniy, a solo tech founder based in Europe with ~20 years of experience in the industry. It all started as a simple idea: build a web server and infrastructure purpose-built for AI services and agents -something with lower latency that could handle thousands of concurrent connections. That's why I chose Linux as the main platform and io_uring as the foundation (I'd worked with it before and knew how fast it could be). From what I could see, existing infrastructure just wasn't ready for AI workloads.So I built a completely new web server from scratch, designed as a foundation for AI agents. Under the hood: Linux + io_uring, Rust, and kTLS. The result is faster than nginx (one of the best out there) with lower latency (1.3-1.8x).
About the website: I'll be honest some of the content was AI-generated, and some was AI-polished. Marketing isn't exactly my strong suit (solo tech founder, remember?). As for Windows/macOS support - it's important, but not the priority right now. The current focus is on polishing the web server, finishing server-sent events (almost there).
