How to Access Your Local Ollama AI from Anywhere Using Tailscale (2026)

Run Ollama on your Mac at home and access it securely from your laptop, phone, or office — no port forwarding, no cloud costs. Here's the exact setup.

chatgpt, claude, gemini, ai, artificial intelligence, chatbot, openai, anthropic, assistant, prompt, llm, ollama, local ai, notebooklm
Reading Tools

Listen & Follow

Hear the article while spoken text is highlighted

00:00
00:00

Quick Answer

Run Ollama on your Mac at home and access it securely from your laptop, phone, or office — no port forwarding, no cloud costs. Here's the exact setup.

  • A Mac running Ollama (your AI server — Mac Mini, MacBook, anything)
  • A second device you want to access from (laptop, phone, tablet)
  • A free Tailscale account (tailscale.com — free for personal use, up to 100 devices)
*As an Amazon Associate I earn from qualifying purchases.
TL;DR: Install Tailscale on your Mac (Ollama server) and your laptop/phone. Set OLLAMA_HOST=0.0.0.0 on the server. Point your client at the Mac’s Tailscale IP on port 11434. Done — your private AI is accessible from anywhere, encrypted, no open ports.

The gap between “I have a local AI server” and “I can use it from anywhere” used to require setting up a VPN, opening firewall ports, or renting a VPS to proxy traffic. Tailscale closes that gap in about 10 minutes. It creates a private encrypted network between your devices — your Mac Mini at home and your laptop at the coffee shop behave as if they’re on the same LAN.

Here’s the exact setup.

“Tailscale + Ollama is the combination that makes a home AI server actually usable. Without remote access, your Mac Mini only helps you when you’re home.”

What You Need

Prerequisites:

  • A Mac running Ollama (your AI server — Mac Mini, MacBook, anything)
  • A second device you want to access from (laptop, phone, tablet)
  • A free Tailscale account (tailscale.com — free for personal use, up to 100 devices)
  • 15 minutes

Step 1: Configure Ollama to Listen on All Interfaces

⚙️

Set OLLAMA_HOST Environment Variable

Makes Ollama accept connections from other devices, not just localhost

By default, Ollama only listens on 127.0.0.1 (your own Mac). To let other devices connect, you need to tell it to listen on all interfaces.

Open Terminal on your Mac server and run:

launchctl setenv OLLAMA_HOST "0.0.0.0"

Then restart Ollama (quit from the menu bar icon and relaunch). For a permanent fix that survives restarts, add it to your shell profile (~/.zshrc):

echo 'export OLLAMA_HOST=0.0.0.0' >> ~/.zshrc && source ~/.zshrc

Step 2: Install Tailscale on Both Devices

🔒

Tailscale Setup

Free, 5 minutes, creates an encrypted private network between your devices

Free for personal use
Mac, iPhone, Windows, Android
WireGuard-based encryption
  1. Go to tailscale.com and create a free account
  2. Download and install Tailscale on your Mac AI server
  3. Sign in on the Mac — it appears in your Tailscale admin panel
  4. Download and install Tailscale on your laptop or phone
  5. Sign in with the same account on all devices
  6. In the Tailscale admin panel at login.tailscale.com/admin/machines, find your Mac’s Tailscale IP — it’ll look like 100.x.x.x

Step 3: Connect and Test

✅

Test Your Connection

From your laptop (anywhere in the world), verify Ollama is reachable

On your laptop, with Tailscale running, open Terminal and run (replace with your Mac’s Tailscale IP):

curl http://100.x.x.x:11434/api/tags

You should see a JSON list of your installed Ollama models. If you do — it works. Your private AI is now accessible from anywhere.

To use it with Open WebUI on your laptop, set the Ollama API base URL to http://100.x.x.x:11434 in Open WebUI settings. Same for any LLM client that accepts a custom API URL.

Step 4: Access from iPhone or iPad

Install Tailscale from the App Store on your iPhone. Sign in. Now apps like Enchanted LLM (a native iOS Ollama client) can connect to your home Mac by pointing at http://100.x.x.x:11434. You have a private AI assistant on your phone that runs on your Mac hardware — no API costs, no data leaving your network.

Security: What You Should Know

Important security notes:

  • Tailscale traffic is encrypted end-to-end using WireGuard — nobody between your devices can intercept it
  • Only devices signed into your Tailscale account can access your Ollama server — it’s not open to the internet
  • Do NOT use OLLAMA_HOST=0.0.0.0 without Tailscale (or another VPN) — you’d be exposing your AI server to your local network and potentially the internet
  • Optionally, add Tailscale ACLs to restrict which devices can reach which ports if you’re on a shared network

Bonus: Use a Stable Hostname Instead of IP

Tailscale’s MagicDNS feature assigns each device a stable hostname — so instead of 100.x.x.x:11434, you can use mac-mini.tail12345.ts.net:11434. Enable MagicDNS in your Tailscale admin panel → DNS settings. Now you never need to remember an IP address again.

Related Content

Run AI Locally on Mac Mini: The Complete Guide

Build Your Own Private ChatGPT with Open WebUI + Ollama

Frequently Asked Questions

Is Tailscale free?

Yes — Tailscale’s personal plan is free for up to 100 devices and 3 users. For a home AI server setup connecting your own Mac, laptop, and phone, you’ll never need to pay.

Does this work on iPhone and Android?

Yes. Tailscale has apps for iOS and Android. Any device signed into your Tailscale account can reach your Ollama server. On iOS, apps like Enchanted LLM support a custom Ollama host URL. On Android, use any OpenAI-compatible client pointed at your Ollama URL.

Will this work if my Mac is on a home router with NAT?

Yes — that’s exactly what Tailscale is designed for. It handles NAT traversal automatically using DERP relay servers as fallback when direct connections aren’t possible. In practice, most connections are direct (peer-to-peer) with relay only as a fallback.

Your private AI, everywhere.

Ten minutes of setup, and your home AI server follows you wherever you go. No API costs. No data leaving your control. This is what local AI is supposed to feel like.

See the Full Local AI Setup Guide →

Subscribe now on Telegram
Next guide coming up
XfWA