n8n integration

Run Luntrex Chat Models from the n8n workflow canvas.

Connect the Luntrex Chat Model in the n8n workflow canvas, point it at your local gateway, and orchestrate multi-model responses with the same dark UI the rest of our docs use.

Configure the Luntrex Chat Model node

This guide walks through updating the node’s base URL, injecting an Luntrex API key, and tuning parameters so n8n workflows can call your router just like OpenAI.

Luntrex Base URL
https://luntrex.com/api/v1/path

Integration steps

Add the node Drag an OpenAI Chat Model node onto your n8n canvas. This node supports custom endpoints, which lets it work seamlessly with Luntrex.
Point to Luntrex Open the node’s credentials panel and replace the default base URL with https://luntrex.com/api/v1/path. This directs requests to your local Luntrex gateway.
Create and paste an API key Generate a key in the Luntrex dashboard, then paste it into the node’s API Key field. Keys can be rotated in the dashboard without editing your workflows.
Adjust parameters if needed In the node parameters, choose the model identifier you want to route (for example, openai/gpt-4o-mini or a custom blueprint). Luntrex will handle fallbacks and guardrails according to your policies.
Run the workflow Save the node, connect it to your trigger, and execute the workflow. Responses, usage, and moderation results will stream back through the node’s output.
Tip: n8n environment variables work great for storing API keys. Set N8N_Luntrex_API_KEY and reference it in the credentials dialog to avoid hardcoding secrets.

Key fields to verify

  • Resource set to Chat with mode Completions.
  • Base URL https://luntrex.com/api/v1/path
  • Model set to the identifier you configured in Luntrex (changeable in node parameters).
  • Temperature / Max tokens aligned with your routing policies.

Troubleshooting

  • If you receive Unauthorized, double-check the API key and ensure it’s active in the dashboard.
  • For connection errors, confirm the Luntrex account is running and accessible from the n8n host.
  • To switch models during testing, edit the node parameter and re-run — no workflow rebuild needed.
Back to top