ollama run is inconvenient — the model should run in the background, survive reboots, and be accessible from a phone or another computer on the network. Below — autostart on Linux and Mac, opening LAN access and a web UI on top of the console.Project card — stars, tech stack, link to sources — in the REDDYX catalog: open-webui/open-webui. The engine itself — ollama/ollama.
AUTOSTART ON LINUX VIA SYSTEMD
The official Ollama installer on Linux already creates a systemd service automatically, but if it is running manually from the terminal for you — move it to systemd:
sudo tee /etc/systemd/system/ollama.service << 'EOF'
[Unit]
Description=Ollama Service
After=network-online.target
[Service]
ExecStart=/usr/local/bin/ollama serve
User=ollama
Restart=always
Environment=OLLAMA_HOST=0.0.0.0:11434
[Install]
WantedBy=multi-user.target
EOF
sudo systemctl daemon-reload
sudo systemctl enable --now ollama
The flag OLLAMA_HOST=0.0.0.0:11434 is key for network access; without it the service listens only on 127.0.0.1 and is inaccessible from other devices.
AUTOSTART ON MAC VIA LAUNCHD
If Ollama is installed as a standalone app — it launches automatically at login via LaunchAgent. For manual binary installation, autostart can be configured with a similar plist file in ~/Library/LaunchAgents/, but most users find it simpler and more reliable to install the official .dmg — autostart is already built in.
ACCESS OVER LOCAL NETWORK FROM PHONE OR ANOTHER COMPUTER
- Make sure the server listens on
0.0.0.0, not onlylocalhost(variableOLLAMA_HOST). - Find the server's local IP:
ip addr(Linux) oripconfig getifaddr en0(Mac). - From a client device on the same Wi‑Fi network, access
http://IP_СЕРВЕРА:11434. - Check the firewall — on Linux port 11434 may be closed by default (
sudo ufw allow 11434).
For internet access (not just the home network), you need a full reverse‑proxy with HTTPS and authentication — exposing the Ollama port directly to the outside without protection is not advisable, as it has no built‑in authorization.
WEB UI INSTEAD OF CONSOLE
Open WebUI provides a full‑featured chat interface on top of Ollama — chat history, model switching, document upload for RAG. It is started with a single command alongside an already running Ollama:
docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -e OLLAMA_BASE_URL=http://host.docker.internal:11434 -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main
After that the interface is available at http://localhost:3000 — the console is no longer needed for everyday work, only for managing models (ollama pull, ollama list).
Frequently Asked Questions
Is it safe to expose Ollama to the network without a password?
Within a trusted home network — acceptable. Exposing to the internet without a proxy with authentication — no, Ollama itself has no built‑in access check.
Will Ollama continue running after a computer reboot?
Yes, if systemd is configured (Linux) or the official app is installed (Mac/Windows) — the service starts automatically on system boot.
Can multiple models be kept loaded in memory at the same time?
Yes, via the OLLAMA_MAX_LOADED_MODELS variable, but each loaded model consumes its own memory — calculate the total volume for all simultaneously active models.
We review fresh open‑source tools every day — from local LLMs to security scanners. If you want to be among the first to see what actually works in 2026 — subscribe to the REDDYX AI Telegram channel — new repositories every 30‑60 minutes.
