A new unofficial CLI and MCP server lets developers and AI agents manage Lambda cloud GPU instances. Written in Rust, the tool provides direct terminal commands for launching, listing, and terminating GPUs. Its standout feature is an MCP server integration, allowing AI assistants like Claude to spin up infrastructure on demand using natural language prompts.

The project addresses the manual friction of managing cloud GPUs for AI development. It connects to Lambda's API to check availability and pricing. Key features include automatic Slack, Discord, or Telegram notifications when instances are SSH-ready, and support for loading API keys via 1Password or other secret managers. The tool is MIT licensed.

While unofficial and unaffiliated with Lambda, this community-built utility fills a niche for developers wanting to automate their GPU workflows. By integrating with the Model Context Protocol, it bridges the gap between AI coding assistants and cloud infrastructure. Users can now ask an agent to provision a powerful H100 instance and run a job, streamlining the entire process.