OpenRouter Chat CompletionInstall

Summary

Send chat completion requests to the OpenRouter API using either a UI-built message array or a complete raw JSON request body, and receive the model's generated response.

Inputs

• Input Mode (required) — Choose between 'Messages' to build messages via UI fields or 'Raw Body' to provide a full JSON request body (needed for multimodal/vision models).
• Model (required) — Select an OpenRouter chat model (required when Input Mode is 'Messages').
• Messages (required) — Provide one or more messages each with a role (system, user, assistant) and content to send to the model (used only when Input Mode is 'Messages').
• Options — Configure generation parameters such as temperature, max tokens, JSON mode, top P, frequency penalty, and presence penalty (only available when Input Mode is 'Messages').
• Request Body (JSON) (required) — Input the complete request body as JSON including at minimum 'model' and 'messages' fields (required when Input Mode is 'Raw Body').

Output shape

a list of JSON objects corresponding to each input item, each containing the model's response content, usage statistics, finish reason, model ID, OpenRouter response ID, and original input data

Responses include error details per item if continuation on failure is enabled. The node throws on blocked models or invalid inputs. The output is paired with the respective input item index.

Links

• n8n community nodes installation guide