Cleric is a self-learning AI site reliability engineer: an agent that investigates production incidents on its own and gets better at it with every incident it sees.
When an incident occurs, Cleric autonomously starts investigating — correlating signals across logs, metrics, and configurations the way an experienced engineer would, rather than reacting to individual alerts. Findings arrive directly in Slack with links to the supporting evidence, and each conclusion carries a confidence score so engineers know how much to trust it. For complex cases, engineers can guide Cleric’s reasoning conversationally or dig into detailed diagnostics in its web interface. What makes Cleric distinctive is the learning loop: it learns from every incident, alert, and human decision, including the feedback engineers give on its investigations, so its signal-to-noise ratio improves the longer it runs in your environment. It works with read-only access to existing observability, CI/CD, and incident tooling such as Datadog and Grafana, so teams do not need to rebuild their stack to adopt it.
Early adopters report reclaiming twenty to thirty percent of the engineering capacity previously lost to repetitive troubleshooting, and production users such as BlaBlaCar have run Cleric since early 2025, with investigations presented transparently enough for engineers to audit the reasoning.
Cleric is sold through its sales team and does not publish list pricing. Teams typically start with a scoped deployment against their existing observability stack.
Cleric is for SRE, platform, and software engineering teams that spend a large share of their on-call time on repetitive troubleshooting and want an agent that compounds what it learns about their systems. It also fits growing teams whose systems are becoming too complex for purely manual on-call.