Rapid Lama

rlama is a lightweight CLI for fast, single-turn prompts against an Ollama server.

What It Does

• Sends one prompt per command execution.
• Uses Ollama's /api/generate endpoint with stream: false.
• Prints the model response directly to stdout.

Requirements

• Node.js 18+
• An Ollama server running locally or remotely
• curl available in your shell

Installation

As a global CLI

npm install -g rapid-lama
rlama "Summarize recursion in one paragraph."

From source (local development)

pnpm install
pnpm build

After building, run:

node dist/cli.js "What is the capital of Italy?"

Usage

rlama [options] <message>

Examples:

rlama "Explain HTTP status code 429"
rlama "Write a short commit message for a bug fix"

Help:

rlama --help

Environment Variables

• OLLAMA_HOST: Ollama base URL. Default: http://localhost:11434
• OLLAMA_MODEL: Model name. Default: llama3.2
• OLLAMA_SYSTEM: Optional system prompt override

Example:

OLLAMA_MODEL=llama3.1 OLLAMA_HOST=http://127.0.0.1:11434 rlama "Give me 3 Linux tips"

Development Scripts

• pnpm build: Compile TypeScript and mark dist/cli.js executable
• pnpm dev: Run dist/cli.js directly

License

MIT