Hi everyone, I'm the founder behind ReliaMesh. I started working on this because traditional infrastructure monitoring doesn't always capture whether AI agents are actually completing their tasks. An API can return HTTP 200 while an agent produces malformed output or starts failing after a model or prompt update. ReliaMesh is my attempt to make those regressions easier to detect and investigate. The current v0.2.0 release includes a Python SDK, self-hosted server, deterministic regression detection, and incident recovery tracking. There's also a reproducible synthetic demonstration in the repository. It's early-stage software, and I'm particularly interested in feedback from engineers operating AI agents in production. GitHub: https://github.com/PrefilerLabs/... Would love to know how others are approaching agent reliability today.

ReliaMesh

Open-source reliability infrastructure for AI agents

Open-source reliability infrastructure for AI agents