CLI reference

alcf-proxy <command> [options]

serve

Start the translating gateway (foreground, resident).

alcf-proxy serve [--host HOST] [--port PORT] [--cluster C] [--framework F] [--model M] [--config PATH]

Flag	Default	Description
--host	127.0.0.1	Bind host. Non-loopback requires ALCF_PROXY_API_KEY.
--port	11445	Bind port.
--cluster	sophia	Primary cluster.
--framework	vllm	Primary framework.
--model	openai/gpt-oss-120b	Default model id.
--config	(search paths)	Explicit config file path.

Exposes: POST /v1/chat/completions, POST /v1/responses, POST /v1/messages, GET /v1/models, GET /health.

models

List models currently served by ALCF (across clusters/frameworks).

alcf-proxy models          # human-readable
alcf-proxy models --json   # machine-readable

health

Print token status (JSON), including hours until expiry.

alcf-proxy health

auth

Manage Globus authentication (see Authentication).

alcf-proxy auth login      # one-time interactive browser login
alcf-proxy auth status     # authenticated? token path + hours left

config

Manage configuration (see Proxy config).

alcf-proxy config init     # write a starter config.yaml
alcf-proxy config show     # print the effective (merged) config

HTTP endpoints (when serving)

Method & path	Purpose
POST /v1/chat/completions	OpenAI Chat (Aider, OpenCode, OpenAI SDK)
POST /v1/responses	OpenAI Responses (Codex)
POST /v1/messages	Anthropic Messages (Claude Code)
GET /v1/models	model list
GET /health	status, active endpoint, token hours left