AI-Powered Operations
4 AI providers, 80+ Kubernetes tools — the AI never touches your cluster directly.
Choose Your AI Provider
Choose the provider that fits your team. Anthropic's Claude for reasoning depth, OpenAI's GPT-4 for breadth, Google's Gemini for long context, or OpenRouter to route across dozens of models from a single key. Switch providers at any time — all share the same 80+ tool set and security constraints: every request goes through the SRExpert backend, never directly to your cluster.
How AI Operations Works
Three layers separate the AI from your cluster, ensuring safe, auditable, read-only operations.
SRE CLI Pod
Deploys as a pod inside your cluster via Helm. Connects to the SRExpert backend API — the AI never touches your Kubernetes API directly.
AI Gateway
The backend validates every AI request against a strict allowlist, executes read-only operations, and returns sanitized results.
Read-Only
No write operations — ever. Scaling, deleting, and editing stay in the UI or kubectl. Every query is audit-logged with full context.
What the SRE CLI Can Do
Not a chatbot bolted onto kubectl. The SRE CLI ships with purpose-built, read-only tools across six domains — the AI picks the right ones and synthesizes results into clear answers.
Workload Analysis
List pods, deployments, services, jobs. View status, replicas, age, and configuration across namespaces.
Monitoring & Metrics
CPU/memory metrics per pod and node, resource trends, recent events, and node readiness checks.
Security Analysis
Ask the AI to run vulnerability scans, check CIS/NSA compliance, analyze RBAC permissions, and detect misconfigurations — all via natural language.
Network Analysis
Network policies, exposed services, traffic flow analysis, and namespace isolation verification.
Storage Management
PVCs, PVs, storage classes — view capacity, usage, and binding status across clusters.
Troubleshooting
Describe resources, search logs, analyze errors, get root cause analysis, and best practice recommendations.
Just Ask
No kubectl flags or PromQL to memorize. Describe what you need — the CLI does the rest.
Security by Architecture
SRExpert addresses AI trust at the architectural level — not with guardrails or prompt engineering, but by physically separating the AI from your cluster. Every interaction is audit-logged and validated.
No Direct Cluster Access
AI communicates only with the SRExpert backend — never with the Kubernetes API or etcd directly.
Read-Only by Design
All 80+ tools are read-only. Write operations require the SRExpert UI or kubectl.
Full Audit Trail
Every query, tool invocation, and response is logged with user, timestamp, and provider.
RBAC-Aware
The CLI respects your existing RBAC policies. Users only see data they are authorized to access.
Documentation
Go deeper in the SRExpert docs:
Start with SRExpert today
Connect your first cluster in 5 minutes. See every workload, enforce security policies, and let AI handle the noise.
Replace your monitoring stack, alerting tools, and compliance spreadsheets with one platform built for Kubernetes teams.
Free tier available — no credit card needed.