Running a local LLM with LM Studio β†’ n8n β€” fixed workflow (real case, step-by-step)


:white_check_mark: Update (New LM Studio Feature β€” LM Link)

In this guide I recommended β€œlimiting LM Studio to your LAN and not exposing it to the public internet”, this is still a good security practice.

However, LM Studio recently introduced :backhand_index_pointing_right: LM Link, which changes the remote-access story in a safer way.

LM Link allows you to access your local models remotely without opening ports or exposing your server publicly. It creates an end-to-end encrypted private network between your devices using a mesh VPN (powered by Tailscale), so your machines communicate securely without being visible on the internet. (:backhand_index_pointing_right: LM Studio)

β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”

Practically, this means:

  • You can keep LM Studio locked down (no public exposure)

  • Still use your models from another device or location

  • n8n can continue connecting via localhost:#### as usual, even when the model runs on another machine


So the original advice remains valid, don’t expose LM Studio publicly, but LM Link is now a secure alternative for remote access instead of port forwarding or public endpoints.

1 Like