CLI Proxy Web icon

CLI Proxy Web

Call CLI Proxy Web API endpoints from n8n

Actions6

Create Response

AI-generated

Summary

Sends the configured model and input to the CLI Proxy Web Responses API at /v1/responses, with optional generation settings and extra request fields. The input accepts a JSON string, array, or object and is required at runtime.

Inputs

  • Model (required) — Model name; defaults to gpt-5-mini.
  • Input (required) — Responses API input as a JSON string, array, or object. Defaults to Summarize the latest request activity.; blank input causes an error.
  • Options → Stream — Whether to request streaming; defaults to false.
  • Options → Reasoning Effort — Proxy-specific reasoning effort: none, low, medium, high, or xhigh. none is omitted from the request.
  • Options → Temperature — Sampling temperature, from 0 to 2; defaults to 1.
  • Options → Maximum Output Tokens — Maximum output token count; defaults to 1024.
  • Options → Text Verbosity — Responses API verbosity: none, low, medium, or high. none is omitted from the request.
  • Options → Metadata — JSON object attached as request metadata; defaults to {}.
  • Options → Additional Body — JSON object merged into the request body. Its fields override matching generated fields.

Output shape

One n8n item per input item, with the API response in json.

The operation uses the required CLI Proxy Web API credentials. If Continue On Fail is enabled, a failed request produces an item with json.error; otherwise the node throws an error.

Links

Discussion