Port 11434

What’s running on port 11434 on my Mac?

11434 is Ollama’s local API. Every app that talks to your local models, from Open WebUI to editor plugins and coding agents, connects to this port.

Commonly used by

  • OllamaThe Ollama app or ollama serve listens on 127.0.0.1:11434.
  • Open WebUI, Msty, Enchanted…Chat front ends that connect to 11434; they’re clients, not the listener.
  • Editor and agent integrationsContinue, Zed and other tools point their local-model setting at http://localhost:11434.

Check what’s on port 11434

lsof -nP -iTCP:11434 -sTCP:LISTEN
curl http://localhost:11434

You’ll see ollama holding the port and Ollama is running from curl. If you also see an Ollama app process and an ollama serve you started yourself, only one can hold 11434: the second fails with address already in use. Quit one of them.

ollama ps          # models loaded right now, and their memory

With Stat Space: statspace port 11434.

Stop or move it

  • Menu bar app: Quit Ollama from its menu bar icon.
  • Terminal: Ctrl-C on ollama serve.

To change the address, set OLLAMA_HOST. GUI apps on macOS don’t read your shell profile, so the app needs launchctl:

# Terminal server
OLLAMA_HOST=127.0.0.1:11500 ollama serve

# Menu bar app (then quit and reopen Ollama)
launchctl setenv OLLAMA_HOST 127.0.0.1:11500

The ollama CLI reads the same variable to find the server, so export it in your shell too.

Security: OLLAMA_HOST=0.0.0.0

Ollama has no authentication. Whoever can reach 11434 can run your models, pull new ones onto your disk, and delete the ones you have. Scans regularly find many thousands of Ollama servers exposed to the internet.

  • The default 127.0.0.1:11434 is local only.
  • OLLAMA_HOST=0.0.0.0 (or the app’s expose-to-network setting, in versions that have one) opens it to every network you join, including café Wi-Fi.
  • If another device needs it, bind to your Mac’s private address on a trusted network, put an authenticated proxy in front, or use a VPN like Tailscale.

Check with lsof: 127.0.0.1:11434 is right, *:11434 is reachable by others. More in dev server exposed to the network.

Seeing it live

Stat Space lists 11434 with the process behind it, how long it’s been up and whether it’s bound to all interfaces. Its Apps view shows how much memory Ollama is holding while a model stays loaded.

Questions

How do I check if Ollama is running?

Run curl http://localhost:11434. It replies Ollama is running. ollama ps lists the models currently loaded in memory.

How do I change Ollama's port?

Set OLLAMA_HOST=127.0.0.1:11500. For ollama serve in Terminal, prefix the command. For the menu bar app, run launchctl setenv OLLAMA_HOST 127.0.0.1:11500 and restart Ollama. Clients need the same address.

Why is Ollama using so much memory when I'm not chatting?

A loaded model stays in memory for 5 minutes after the last request by default. ollama stop <model> unloads it now, and OLLAMA_KEEP_ALIVE changes the timeout.

Get started

Know what your Mac is doing, without adding to it.

Free, no account. Pro is $9.99 once for the first 100 buyers.