Open-source systems for AI-native operations, observability, automation, security, and cost governance.
XOps Labs is a place for practical infrastructure: tools that help teams run modern software and AI workloads with clearer signals, safer defaults, and less operational guesswork.
flowchart LR
Build["Build"] --> Observe["Observe"]
Observe --> Govern["Govern"]
Govern --> Automate["Automate"]
Automate --> Secure["Secure"]
Secure --> Improve["Improve"]
Improve --> Build
XOps Labs is not a single-project org. It is an open-source lab for tools that sit close to production operations: the layer where platform engineering, observability, security, FinOps, and AI systems all start to overlap.
| Workstream | Direction |
|---|---|
| AI-native operations | LLMOps, model usage visibility, provider telemetry, agent-aware infrastructure |
| Observability | Prometheus metrics, Grafana dashboards, OpenTelemetry, health signals, alert-ready data |
| FinOps and governance | Cost attribution, budget signals, FOCUS-style records, showback and chargeback workflows |
| Platform automation | Kubernetes, containers, CI/CD, release workflows, deployment scaffolding, day-2 operations |
| Security operations | Least privilege, secret-safe patterns, CodeQL, supply-chain metadata, secure defaults |
| Developer experience | Practical docs, local-first demos, repeatable examples, boringly useful tooling |
The first public XOps Labs project: a self-hosted Prometheus and OpenTelemetry exporter for LLM usage, token volume, request counts, prompt caching, and cost telemetry across major AI providers.
It is one example of the broader XOps Labs direction: make invisible operational signals visible, keep the system self-hosted, and give teams data they can actually act on.
More projects will follow the same pattern: focused tools, open standards, production-minded defaults, and clear docs.
- Open source by default
- Self-hosted before SaaS-dependent
- Open standards over proprietary lock-in
- Low-cardinality, production-safe telemetry
- Security and cost treated as operational concerns, not afterthoughts
- Tools should be easy to run locally and credible in production
Bring issues, ideas, provider integrations, dashboards, examples, docs, tests, and real-world operational feedback. XOps Labs is built around practical tools that become sharper when operators use them.
Explore the repositories: github.com/xops-labs
