Actions5
Compact (LLM Directo)
AI-generatedSummary
Calls Orbital’s compact endpoint for a direct LLM request without running an agent. Choose Raw mode to send system and user messages, or Prompt mode to run a saved prompt template with variables; select a model and set the maximum output tokens.
Inputs
- Autenticacion (required) — Authenticate with a Bearer API key or job JWT in Token, or select a saved n8n Header Auth credential.
- Modo — Choose Raw (default) to provide messages directly, or Prompt to use a saved template.
- Modelo — LLM to call; defaults to GPT-4o Mini. Available choices include GPT-4o, GPT-4.1 Mini, Claude Sonnet 4, Claude Haiku 3.5, Gemini 2.0 Flash, and Llama 3.1 8B.
- System Prompt — Optional developer message included in Raw mode when non-empty.
- User Message (required) — Required user message for Raw mode.
- Prompt ID — Saved prompt template ID used in Prompt mode.
- Prompt Variables — Optional name/value entries passed to the saved prompt in Prompt mode.
- Max Tokens — Maximum output tokens, default 1000.
Output shape
A single n8n item containing the response data.
The JSON includes operation: "compact" and the fields returned by Orbital. Request failures are returned in the item as success: false with error details.