Chat with local AI privately. No cloud. No subscription. Runs on your own Ollama server.
macOS 10.12 or later. Signed and notarized — opens without any warnings.
Unzip and run Zikporx AI.exe. Windows may show a SmartScreen warning — click More info → Run anyway.
Unzip and run ./Zikporx AI. You may need to run chmod +x Zikporx AI first.
Zikporx AI Desktop connects directly to Ollama — an open-source tool that runs AI models locally on your computer.
Download from ollama.com/download and install it. It runs silently in the menu bar.
Open Terminal and run:
ollama pull llama3.2:3b
Swap llama3.2:3b for any model at ollama.com/library. For image analysis use llava.
Launch Zikporx AI. The default address http://localhost:11434 connects automatically. Open Settings → Test connection to confirm.
Download and install Ollama: ollama.com/download
Open Terminal on that machine:
ollama pull llama3.2:3b
By default Ollama only listens on 127.0.0.1. Start it with:
OLLAMA_HOST=0.0.0.0 ollama serve
To make this permanent on Linux (Ubuntu / Debian):
sudo systemctl edit ollama
In the editor, add:
[Service] Environment="OLLAMA_HOST=0.0.0.0"
Then apply:
sudo systemctl daemon-reload && sudo systemctl restart ollama
Allow port 11434 on the Ollama machine:
sudo ufw allow 11434
On the Ollama machine run:
ip addr show | grep "inet " | grep -v 127.0.0.1
On macOS: ipconfig getifaddr en0 — you'll get something like 192.168.1.100.
Open Zikporx AI → Settings → set the server address to your Ollama machine's IP:
http://192.168.1.100:11434
Tap Test connection — green "Connected" means you're ready.
Ollama isn't running or isn't reachable. Make sure you started it with OLLAMA_HOST=0.0.0.0 ollama serve.
Check the firewall (sudo ufw allow 11434) and confirm both computers are on the same network.
The model name in Settings must match exactly what ollama list shows. Names are case-sensitive.
Large models (7B+) need a capable GPU or CPU. Try llama3.2:3b or phi3:mini for faster responses.