Ollama CLI - Run LLMs Locally from the Command Line

		# Download and install from ollama.com
# Or use Homebrew
brew install ollama
	

		curl -fsSL https://ollama.com/install.sh | sh
	

		ollama --version
	

		# Download and run Llama 3.2 (3B parameters)
ollama run llama3.2
 
# Or try Mistral
ollama run mistral
 
# Or Phi-3 (smaller, faster)
ollama run phi3
	

		# Interactive mode
ollama run llama3.2
 
# Single query
ollama run llama3.2 "Explain quantum computing in simple terms"
 
# With system prompt
ollama run llama3.2 "You are a helpful coding assistant" "Write a Python function to reverse a string"
	

		# See locally installed models
ollama list
 
# Output shows:
# NAME            SIZE    MODIFIED
# llama3.2:latest 2.0GB   2 days ago
# mistral:latest  4.1GB   1 week ago
	

		# Pull a model without running
ollama pull llama3.2
 
# Pull specific version/size
ollama pull llama3.2:3b
ollama pull llama3.2:7b
 
# Pull vision model
ollama pull llava
	

		# Remove a model to free space
ollama rm llama3.2
 
# Remove specific version
ollama rm llama3.2:7b
	

		# Run a vision model
ollama run llava
 
# In the interactive prompt
>>> What's in this image? /path/to/image.jpg
	

		ollama run llava "Describe this image" --image /path/to/image.jpg
	

		# Set personality/behavior
ollama run llama3.2 
  --system "You are a pirate. Respond to all queries in pirate speak."
	

		# Adjust temperature (creativity)
ollama run llama3.2 --temperature 0.8
 
# Set context window
ollama run llama3.2 --context-length 4096
 
# Limit tokens
ollama run llama3.2 --num-predict 100
	

		# Create Modelfile
cat > Modelfile << EOF
FROM llama3.2
 
# Set temperature
PARAMETER temperature 0.7
 
# Set system prompt
SYSTEM You are a helpful AI assistant specialized in web development.
 
# Set context length
PARAMETER num_ctx 4096
EOF
 
# Build custom model
ollama create webdev-assistant -f Modelfile
 
# Run it
ollama run webdev-assistant
	

		# Server starts automatically with ollama run
# Or start explicitly
ollama serve
	

		curl http://localhost:11434/api/generate -d '{
  "model": "llama3.2",
  "prompt": "Why is the sky blue?",
  "stream": false
}'
	

		curl http://localhost:11434/api/chat -d '{
  "model": "llama3.2",
  "messages": [
    {
      "role": "user",
      "content": "What is recursion?"
    }
  ],
  "stream": false
}'
	

		curl http://localhost:11434/api/tags
	

		curl http://localhost:11434/api/generate -d '{
  "model": "llama3.2",
  "prompt": "Write a story",
  "stream": true
}'
	

		import requests
import json
 
def chat(prompt):
    response = requests.post(
        'http://localhost:11434/api/generate',
        json={
            'model': 'llama3.2',
            'prompt': prompt,
            'stream': False
        }
    )
    return response.json()['response']
 
result = chat("Explain async/await in Python")
print(result)
	

		#!/bin/bash
 
# AI code reviewer
ai_review() {
    local code="$1"
    curl -s http://localhost:11434/api/generate -d "{
        "model": "codellama",
        "prompt": "Review this code:\n${code}",
        "stream": false
    }" | jq -r '.response'
}
 
# Usage
ai_review "$(cat script.py)"
	

		const axios = require('axios');
 
async function generate(prompt) {
    const response = await axios.post('http://localhost:11434/api/generate', {
        model: 'llama3.2',
        prompt: prompt,
        stream: false
    });
    return response.data.response;
}
 
generate('Explain promises in JavaScript').then(console.log);
	

		# Check GPU usage
nvidia-smi
 
# Force CPU only
OLLAMA_NUM_GPU=0 ollama run llama3.2
	

		# Limit concurrent models
OLLAMA_MAX_LOADED_MODELS=1 ollama serve
 
# Set keep-alive duration
ollama run llama3.2 --keep-alive 5m
	

		# Create coding assistant
ollama create coder -f - << EOF
FROM codellama
SYSTEM You are an expert programmer. Provide clear, concise code examples.
PARAMETER temperature 0.3
EOF
 
ollama run coder "Write a React hook for fetching data"
	

		# Summarize documents
find ./docs -name "*.txt" -exec sh -c 
  'ollama run llama3.2 "Summarize: $(cat {})" > {}.summary' ;
	

		# Git commit message generator
git-ai-commit() {
    local diff=$(git diff --staged)
    ollama run llama3.2 "Write a concise git commit message for: $diff"
}
	

		# Install Ollama plugin in Obsidian
# Configure to use local endpoint: http://localhost:11434
	

		# Check disk space
df -h
 
# Clear cache
rm -rf ~/.ollama/models
 
# Retry pull
ollama pull llama3.2
	

		# Use smaller model
ollama run phi3  # 3.8B instead of 7B+
 
# Reduce context length
ollama run llama3.2 --context-length 2048
	

		# Check if server is running
ps aux | grep ollama
 
# Restart server
killall ollama
ollama serve