Relvy AI agent autonomously investigates on-call alerts
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Relvy is a specialized AI agent that autonomously investigates on-call alerts using purpose-built tools for production telemetry and code, producing auditable investigation notebooks teams can review and trust.
- Relvy claims 70% of alerts are resolved in under five minutes via standardized incident response, positioning the product as giving teams "your most tenured engineer on-call all the time."
- Y Combinator-backed Relvy integrates with existing observability stacks — logs, metrics, traces, events, deployments, code analysis, and internal APIs/MCP tools — and works with PagerDuty, importing existing runbooks or generating new ones.
- Relvy recently improved Claude's root-cause-analysis accuracy by 12 percentage points on the OpenRCA benchmark, per a linked company blog post.
- Aniket Dash, Platform Infrastructure Lead at Blend, endorsed Relvy for executing runbooks across Blend's observability stack, citing reduced context switching and operational standardization.
- Relvy is SOC 2 Type II certified with self-host options available, and the company offers Forward Deployed Engineers to tune the agent for customer stacks.
Why it matters: For on-call engineering teams, Relvy's 70%-in-five-minutes claim — backed by a Blend testimonial and SOC 2 Type II certification with self-host options — points to a shift from manual alert triage to reviewing AI-generated notebooks, targeting enterprise customers where compliance and data residency already gate adoption.
