“I built Latch‑X because the tool I needed simply didn’t exist. I started with analytical models, spreadsheets, and traditional tools like Fault Tree Analysis (FTA) or Reliability Block Diagrams (RBD), but they couldn’t capture the dynamic behavior of modern systems—or they ran into complexity limits, like those in Markov models.

That led me to explore Bayesian Networks (BN). I reused logic gates from FTA—such as and:, or:, and n_of_k:—and introduced a new kind of dependency: do:, to better express causality.

The hardest part was modeling failover dynamics in a static BN. After a few sleepless nights and many iterations, I came up with a new probabilistic abstraction: the latch node. It solved that problem—and gave Latch‑X its name.

Once I had the core working, I decided to share it as a SaaS platform so others could benefit from it instantly. Latch‑X started as a declarative personal script, and I chose to keep that declarative structure. But to make it more approachable, I added a simple graphical UI and an AI assistant to help non-technical users as well.

I believe Latch‑X can be used even without deep reliability engineering knowledge. It’s meant to be accessible to architects, engineers, and anyone designing for availability, risk, or downtime.”

Latch‑X is evolving. If you use Latch‑X, I’d really appreciate your feedback: what’s working, what’s not, what’s missing.

— Peter Kováč, Founder & CTO