# Local-first AI DevOps

Local-first AI DevOps is an operating model for infrastructure work where live context, AI reasoning, and operator approvals stay close to the machine already trusted to talk to cloud and cluster APIs.

## What it is

- A context and action workspace around live provider, Kubernetes, repo, and cost evidence.
- A local-first trust boundary for credentials, kubeconfigs, and BYOK model keys.
- A review-before-apply workflow instead of silent hosted automation.

## What it is not

- Not a full observability backend.
- Not a pager or on-call routing system.
- Not a remote IaC orchestrator.
- Not a single-purpose cluster admin UI.

## Hosted vs local

- Local-first keeps credentials and approvals close to the operator.
- Hosted models trade setup simplicity for another trust boundary.
- The architecture choice affects BYOK pricing, compliance review, and how easily AI can be grounded in live evidence.

## Practical implementation

Clanker Cloud is the practical implementation of this category: local runtime, live multi-provider context, reviewed plans, and explicit operator approval before infrastructure changes.

## Current and upcoming support

- Current support includes AWS, GCP, Azure, Kubernetes, Cloudflare, Hetzner, DigitalOcean, Vercel, GitHub, and BYOK model providers.
- Coming next: Ansible and Slurm.
- Ansible and Slurm are planned support, not current GA coverage on this page.

## Related pages

- [What is Clanker Cloud?](https://clankercloud.ai/what-is-clanker-cloud)
- [How Clanker Cloud works](https://clankercloud.ai/how-clanker-cloud-works)
- [Clanker Cloud vs Datadog](https://clankercloud.ai/clanker-cloud-vs-datadog)
- [Clanker Cloud vs PagerDuty](https://clankercloud.ai/clanker-cloud-vs-pagerduty)
- [Local-first AI DevOps vs hosted AI DevOps](https://clankercloud.ai/local-first-ai-devops-vs-hosted-ai-devops)