Hey everyone! I'm Phil, the builder behind Reflex. I've been managing production servers for years — both my own SaaS projects and at my day job. The pattern was always the same: something breaks at 2am, you get the alert, SSH in half-asleep, run the same diagnostic commands you've run a hundred times, apply the same fix, go back to bed. Repeat next week. I built Reflex because most server incidents aren't novel. They're the same handful of failures cycling on repeat — a queue worker dies, PHP-FPM runs out of children, a deploy leaves bad permissions, memory creeps up until OOM kills something important. These are fixable with known steps. So why are we still doing it manually? Reflex puts a lightweight agent on your server that watches for these patterns and runs repair playbooks automatically. The Brain layer uses LLM diagnostics to figure out what's wrong and what to do about it. You can start in manual-confirm mode where it tells you the problem and recommends a fix, and you approve it. Once you trust it, flip it to auto-repair and sleep through the night. It supports PHP, Laravel, Node.js, Python, Django, FastAPI, React, Vue, Next.js, Nuxt, Ruby, Rails, Go, Java, and Spring Boot — basically if it runs on Linux, Reflex can look after it. Would love your feedback. What's the server issue that keeps waking you up?

Reflex

Infrastructure that reacts before you do.

Infrastructure that reacts before you do.