n8n
Discover all workflows in an n8n instance, including the model each workflow uses.
Airia discovers all workflows in an n8n instance as agents, including the model each workflow's LLM node uses. Works with n8n Cloud (including trial) and self-hosted instances.
Authentication model
An n8n API key plus the instance URL.
Customer-side setup steps
- In n8n, go to Settings → n8n API → Create API key and copy the key. (On n8n Cloud the
key is a long JWT beginning with
eyJ; that is expected.) - Copy your instance root URL (for example
https://yourteam.app.n8n.cloud). - In Airia, go to Discover → Connections → n8n and create a new connection with the URL and API key.
Form fields the Airia UI asks for
| Field | Required | Notes |
|---|---|---|
| URL | Yes | Instance root URL; no /api/v1 suffix needed |
| API key | Yes | Stored encrypted, masked on edit |
Why a discovered workflow can show no model
The model shown in the inventory is extracted from the workflow's LLM node. A workflow shows an empty model (and a "Not Supported" proxy status) in three cases:
- The workflow has no LLM node at all (for example, a trigger plus non-AI nodes).
- The LLM node type is not yet recognized for model extraction. OpenAI and Anthropic chat-model nodes are recognized; some other providers' nodes may not be.
- The model parameter is not explicitly set in the node. Stock n8n templates often ship with the model unset; open the node, explicitly select a model, and save.
After fixing the workflow in n8n, refresh the connection in Airia to update the inventory.
Multiple environments
Multiple n8n connections per tenant are supported: one per instance.
Was this page helpful?