Terradev

This is infrastructure as code for GPU compute, giving Claude direct access to provision machines, manage training jobs, and handle inference deployments. With 192 tools in the kit, you're looking at granular control over the full ML pipeline rather than simple API wrappers. Reach for this when you want Claude to spin up GPU clusters, orchestrate distributed training runs, or manage model serving infrastructure without bouncing between CLI tools and cloud consoles. It's built for developers running serious ML workloads who want conversational control over their compute stack.