Return to Home
Ollama Logo

Ollama Cheat Sheet

Your quick guide to Ollama commands and advanced usage.

Model Management

ollama pull <model_name>

Downloads a model from the Ollama library.

Example:

ollama pull llama3
ollama list

Lists all models that you have downloaded.

Example:

ollama list
ollama rm <model_name>

Deletes a model from your local machine.

Example:

ollama rm llama3
ollama cp <source_model> <new_name>

Creates a copy of a model.

Example:

ollama cp llama3 my-llama3-copy

Running Models

ollama run <model_name>

Starts a conversation with a model.

Example:

ollama run llama3
ollama run <model_name> "Your prompt"

Runs a model with a single prompt and exits.

Example:

ollama run llama3 "What is the capital of France?"
/set verbose

Inside a chat, toggles verbose mode to see more details.

Example:

/set verbose
/show info

Inside a chat, shows information about the current model.

Example:

/show info

Modelfile Commands

ollama create <model_name> -f ./Modelfile

Creates a model from a Modelfile.

Example:

ollama create my-custom-model -f ./Modelfile
FROM <base_model>

(In Modelfile) Specifies the base model to use.

Example:

FROM llama3
PARAMETER <name> <value>

(In Modelfile) Sets a parameter for the model.

Example:

PARAMETER temperature 0.7
SYSTEM """..."""

(In Modelfile) Sets a system-level message.

Example:

SYSTEM """You are a helpful AI assistant."""

API & Server

ollama serve

Starts the Ollama server. It usually runs in the background.

Example:

ollama serve
curl http://localhost:11434/api/generate -d '{ ... }'

Use the API to generate a response. (Check docs for JSON).

Example:

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

Fetches the list of local models via the API.

Example:

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

Advanced Tips & Useful Tools

Tool Calling / Internet Access

Models can't access the internet directly. Use "Tool Calling" with a library like ollama-python to have the model request external data (e.g., search results), which your code provides.

Example:

// Your Python code detects this request from the model // then calls a search API and feeds results back in.
Open WebUI (Docker)

A powerful, self-hosted web UI for Ollama. Supports RAG, multi-model chat, and more. Run with Docker for an easy setup.

Example:

docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
Important Websites

Official resources for documentation, updates, community tools, and SOPHIE's Daddy Blog.