You ran ollama run or tried to connect to the API and got:

Error: could not connect to ollama app, is it running?

Or:

connection refused: localhost:11434

This means the Ollama server isn’t running or isn’t accessible. Here are the fixes.

Fix 1: Start the Ollama service

The most common cause — Ollama isn’t running:

# macOS: start the app
open -a Ollama

# Linux: start the service
sudo systemctl start ollama

# Or run directly
ollama serve

Verify it’s running:

curl http://localhost:11434/api/version
# Should return: {"version":"0.x.x"}

Fix 2: Check the port

Something else might be using port 11434:

# Check what's on port 11434
lsof -i :11434  # macOS/Linux
netstat -tlnp | grep 11434  # Linux

# If another process is using it, change Ollama's port
OLLAMA_HOST=0.0.0.0:11435 ollama serve

Fix 3: Docker networking

If Ollama runs in Docker and your app can’t reach it:

# docker-compose.yml
services:
  ollama:
    image: ollama/ollama
    ports:
      - "11434:11434"  # Expose to host

  your-app:
    build: .
    environment:
      # Use service name, not localhost
      - OLLAMA_HOST=http://ollama:11434

Inside Docker, localhost refers to the container itself, not the host. Use the service name (ollama) or host.docker.internal to reach the Ollama container.

# From another container
curl http://ollama:11434/api/version

# From the host machine
curl http://localhost:11434/api/version

Fix 4: Firewall blocking

On Linux, the firewall might block port 11434:

# UFW
sudo ufw allow 11434

# firewalld
sudo firewall-cmd --add-port=11434/tcp --permanent
sudo firewall-cmd --reload

Fix 5: Bind to all interfaces

By default, Ollama only listens on localhost. To access from other machines:

# Listen on all interfaces
OLLAMA_HOST=0.0.0.0:11434 ollama serve

# Or set permanently
# Linux: edit /etc/systemd/system/ollama.service
# Add: Environment="OLLAMA_HOST=0.0.0.0:11434"
sudo systemctl daemon-reload
sudo systemctl restart ollama

Security warning: Only do this on trusted networks. Exposing Ollama to the internet without authentication is a security risk. See our sandbox guide for securing local AI.

Fix 6: Ollama crashed silently

Check the logs:

# macOS
cat ~/.ollama/logs/server.log

# Linux (systemd)
journalctl -u ollama -n 50

# Docker
docker logs ollama

Common crash causes:

• Out of memory (see our OOM fix)
• Corrupted model files (ollama rm model-name and re-pull)
• GPU driver issues (update NVIDIA drivers)

Fix 7: WSL2 on Windows

If using Ollama in WSL2:

# Ollama in WSL2 binds to the WSL IP, not Windows localhost
# Find the WSL IP
hostname -I

# Connect from Windows using that IP
curl http://172.x.x.x:11434/api/version

Or install the native Windows version of Ollama instead of the WSL2 version.

Related: Ollama Complete Guide · Ollama Troubleshooting Guide · Ollama Out of Memory Fix · Ollama Model Not Found Fix · How to Sandbox Local AI Models · How to Set Up Open WebUI