I built a native Torrix node for n8n that logs every LLM call automatically. No code, no webhooks.
Drop the node into any workflow and you get:
-
Tokens and cost per step
-
Full prompt traces
-
End-to-end latency for the full run
-
Budget alerts before you hit your limit
Here is a support triage workflow as an example: install/demos/n8n at main · torrix-ai/install · GitHub
Self-hosted, one Docker command, free forever.
Happy to help anyone set it up.
