Lambda CLI - MCP Server for Lambda Cloud GPU Instances

TL;DR

TL;DR: Lambda CLI is an MIT-licensed community tool that wraps the Lambda cloud GPU API into both a terminal commands and an MCP server, letting AI assistants like Claude and Cursor manage your GPU instances directly.

Source and Accuracy Notes

• Project page: github.com/Strand-AI/lambda-cli
• Source repository: github.com/Strand-AI/lambda-cli (MIT license, verified via license.spdx_id)
• npm package: npmjs.com/package/@strand-ai/lambda-mcp
• HN launch thread: news.ycombinator.com/item?id=46621786
• MCP compatibility badge: modelcontextprotocol.io
• Source last checked: 2026-06-22 (commit verified via GitHub API)

What Is Lambda CLI?

Lambda CLI is an unofficial community-built CLI and MCP server for Lambda, a cloud GPU provider. It is not affiliated with or endorsed by Lambda — but it wraps their cloud API into a ergonomic tool.

The project exposes two interfaces:

• lambda — Direct terminal commands for provisioning, listing, and managing GPU instances
• lambda-mcp — An MCP server that AI coding assistants (Claude via Cursor, VS Code) can use to manage GPU infrastructure conversationally

   ┌─────────────────┐     ┌──────────────────┐
   │  Cursor / VS Code│     │  Lambda cloud GPU  │
   │  (Claude, etc.)  │────▶│  instances         │
   └─────────────────┘     └──────────────────┘
            │                      ▲
            │ MCP protocol         │
            ▼                      │
   ┌─────────────────┐             │
   │  lambda-mcp     │─────────────┘
   │  (your machine) │
   └─────────────────┘

Setup

Prerequisites

• Node.js 18+ (for the npm package)
• A Lambda cloud GPU account and an API key from the Lambda dashboard
• Homebrew (macOS/Linux) or Rust/Cargo (if building from source)

Install via Homebrew

brew install strand-ai/tap/lambda-cli

Install as MCP server (for Cursor/VS Code)

# Cursor
cursor://anysphere.cursor-deeplink/mcp/install?name=lambda&config=%7B%22command%22%3A%22npx%22%2C%22args%22%3A%5B%22-y%22%2C%22%40strand-ai%2Flambda-mcp%22%5D%7D

# VS Code
vscode.dev/redirect/mcp/install?name=lambda&config=%7B%22command%22%3A%22npx%22%2C%22args%22%3A%5B%22-y%22%2C%22%40strand-ai%2Flambda-mcp%22%5D%7D

Or via npx directly:

npx -y @strand-ai/lambda-mcp

Authenticate

Export your Lambda API key:

export LAMBDA_API_KEY=your_key_here

Or use a secret manager command (e.g., 1Password):

export LAMBDA_API_KEY_COMMAND="op read op://Personal/Lambda/api-key"

Optional: Notifications

Get Slack, Discord, or Telegram alerts when your instance is ready:

export LAMBDA_NOTIFY_SLACK_WEBHOOK="https://hooks.slack.com/services/T00/B00/XXX"
export LAMBDA_NOTIFY_DISCORD_WEBHOOK="https://discord.com/api/webhooks/123/abc"
export LAMBDA_NOTIFY_TELEGRAM_BOT_TOKEN="your_bot_token"
export LAMBDA_NOTIFY_TELEGRAM_CHAT_ID="123456789"

Deeper Analysis

CLI Commands

The CLI supports direct commands for managing the full lifecycle of Lambda GPU instances. Key commands include provisioning new instances, listing active instances, checking status, and terminating resources. The exact command surface is defined by the Lambda cloud API — the CLI wraps these into human-readable terminal output.

MCP Server

The MCP server exposes Lambda’s GPU instance management as MCP tools. This means any MCP-compatible AI assistant can call tools like lambda_list_instances, lambda_create_instance, and lambda_delete_instance with natural language. The MCP server handles authentication via the same LAMBDA_API_KEY environment variable.

Limitations

• Unofficial — this is a community project, not a Lambda product. Use at your own discretion.
• Rate limiting and API stability depend on Lambda’s upstream API.
• No warranty — MIT license disclaims liability.

Practical Evaluation Checklist

• [ ] API key obtained from Lambda dashboard
• [ ] Homebrew installation completed successfully
• [ ] CLI authentication verified (e.g., lambda command returns help)
• [ ] MCP server starts without errors
• [ ] Cursor or VS Code MCP integration confirmed
• [ ] Notification webhooks tested (if configured)

Security Notes

• API key — store LAMBDA_API_KEY in environment or a secret manager. Never hardcode it in scripts or commit it to git.
• Webhook URLs — Slack/Discord/Telegram webhook URLs are secrets. Treat them the same as API keys.
• Unofficial project — the tool reads and writes Lambda cloud resources under your account. Review the source code before granting AI assistants access to it.

FAQ

Q: Is this affiliated with Lambda? A: No. Lambda CLI is a community-built, unofficial tool. The README includes a caution banner noting it is not affiliated with or endorsed by Lambda.

Q: Can I use this in production? A: The tool wraps Lambda’s own cloud API, so it is as reliable as Lambda’s API. MIT licensed, but use judgment — for critical infrastructure, prefer Lambda’s official SDKs.

Q: Does it support all Lambda instance types? A: It wraps the Lambda cloud API, which covers the GPU instance types Lambda offers. Check the Lambda dashboard for the full list of available instance families.

Q: What MCP clients are supported? A: Any MCP-compatible client — Cursor, VS Code with MCP extension, Claude Code, and any other tool that implements the Model Context Protocol spec.

Conclusion

Lambda CLI fills a gap for developers who want to manage Lambda cloud GPU resources from the terminal or through AI assistants. The MCP server integration is the standout feature — it lets you provision or tear down GPU instances via natural language in Cursor or VS Code, which is genuinely useful for developers building and iterating on LLM applications.

The MIT license and public GitHub repo make it auditable. The main caveat is that it is unofficial, so keep an eye on the repo for updates when Lambda’s API evolves.

• Repo: github.com/Strand-AI/lambda-cli
• npm: npmjs.com/package/@strand-ai/lambda-mcp