models

Synopsis

ramalama models [options]

Description

List model identifiers exposed by a running inference server, such as one started with ramalama serve. The command queries the server's OpenAI-compatible /v1/models endpoint first (works with llama.cpp, MLX, and other OpenAI-compatible servers), then falls back to the llama.cpp native /models endpoint.

This differs from ramalama list, which shows models downloaded to local storage.

Options

--api-key

OpenAI-compatible API key. Can also be set in ramalama.conf or via the RAMALAMA_API_KEY environment variable.

--help, -h

show this help message and exit

--json

print model list in json format

--url=URL

model server URL (default: http://127.0.0.1:8080)

Examples

List models served on the default local endpoint

$ ramalama models
tinyllama

List models from a specific server in JSON format

$ ramalama models --url http://localhost:1234 --json
["granite3-moe", "tinyllama"]

See Also

ramalama(1), ramalama-serve(1), ramalama-list(1)

---

Jul 2026, Originally compiled by Dan Walsh <dwalsh@redhat.com>