Ollama Commands and Settings: Model Management, Deleting Models, and Environment Variables

# Install on Linux; re-run the same script to upgrade
curl -fsSL https://ollama.com/install.sh | sh

# Show the version
ollama -v

# Start the server manually (listens on 127.0.0.1:11434 by default)
ollama serve

# Download a model
ollama pull llama3.2

# Interactive chat; type /bye to exit
ollama run llama3.2

# One-shot prompt, handy in scripts
ollama run llama3.2 "Explain RAG in three sentences."

# Show timings and token speed
ollama run llama3.2 --verbose

# Control how long the model stays in memory
ollama run llama3.2 --keepalive 30m

# List downloaded models (ls is an alias of list)
ollama list
ollama ls

# Show loaded models, their size, and GPU/CPU split
ollama ps

# Unload a model to free RAM / VRAM (files stay on disk)
ollama stop llama3.2

ollama show llama3.2
ollama show llama3.2 --modelfile   # Modelfile
ollama show llama3.2 --parameters  # parameters
ollama show llama3.2 --template    # template
ollama show llama3.2 --system      # system message
ollama show llama3.2 --license     # license

# Check model names and sizes first
ollama list

# Delete one model
ollama rm llama3.2

# Delete several models at once
ollama rm llama3.2 mymodel

# Disk usage of the model directory (macOS)
du -sh ~/.ollama/models

# Linux (official install script)
sudo du -sh /usr/share/ollama/.ollama/models

# Modelfile
FROM llama3.2
PARAMETER temperature 0.7
PARAMETER num_ctx 8192
SYSTEM You are a concise technical assistant.

# Create the model (-f defaults to Modelfile in the current directory)
ollama create my-assistant -f ./Modelfile

# Run the custom model
ollama run my-assistant

# Copy a model before changing it so the original stays intact
ollama cp my-assistant my-assistant-v2

OLLAMA_HOST=127.0.0.1:11434 OLLAMA_CONTEXT_LENGTH=8192 ollama serve

launchctl setenv OLLAMA_MODELS "/Volumes/Data/ollama/models"
launchctl setenv OLLAMA_KEEP_ALIVE "30m"
# Quit and reopen the Ollama app afterwards

sudo systemctl edit ollama.service

[Service]
Environment="OLLAMA_MODELS=/data/ollama/models"
Environment="OLLAMA_KEEP_ALIVE=30m"
Environment="OLLAMA_NUM_PARALLEL=2"

# A custom model directory must be readable and writable by the ollama user
sudo chown -R ollama:ollama /data/ollama/models

# Reload and restart the service
sudo systemctl daemon-reload
sudo systemctl restart ollama

# View logs
journalctl -e -u ollama

# Listen on all interfaces (set it per platform as shown above)
OLLAMA_HOST=0.0.0.0:11434

OLLAMA_ORIGINS=chrome-extension://*,moz-extension://*,safari-web-extension://* ollama serve

# Server version
curl http://localhost:11434/api/version

# List downloaded models (same as ollama list)
curl http://localhost:11434/api/tags

# List loaded models (same as ollama ps)
curl http://localhost:11434/api/ps

# One-shot generation
curl http://localhost:11434/api/generate -d '{
  "model": "llama3.2",
  "prompt": "Describe Ollama in one sentence.",
  "stream": false
}'

# Chat with an explicit context length
curl http://localhost:11434/api/chat -d '{
  "model": "llama3.2",
  "messages": [{"role": "user", "content": "Why is the sky blue?"}],
  "stream": false,
  "options": {"num_ctx": 8192}
}'

# Preload a model and keep it in memory
curl http://localhost:11434/api/generate -d '{"model": "llama3.2", "keep_alive": -1}'

# Unload a model immediately
curl http://localhost:11434/api/generate -d '{"model": "llama3.2", "keep_alive": 0}'

# ⚠️ Delete a model (same as ollama rm, irreversible)
curl -X DELETE http://localhost:11434/api/delete -d '{"model": "llama3.2"}'

import requests

def chat_with_ollama(prompt, model="llama3.2"):
    response = requests.post(
        "http://localhost:11434/api/chat",
        json={
            "model": model,
            "messages": [{"role": "user", "content": prompt}],
            "stream": False,
        },
        timeout=120,
    )
    response.raise_for_status()
    return response.json()["message"]["content"]

print(chat_with_ollama("Summarize the following text in three bullet points: ..."))