Ollama CLI Commands (Run in PowerShell)
These are the commands you type at the normal PS C:\Users\kadir> prompt to manage models and the Ollama service itself.
| Command | What it does |
|---|---|
ollama --help | Shows every available command. This is the best place to start if you are ever unsure. |
ollama list (or ollama ls) | Lists all models you have downloaded locally. |
ollama pull <model> | Downloads a model from the Ollama library without running it. |
ollama rm <model> | Removes a downloaded model to free up disk space. |
ollama ps | Shows which models are currently running in memory. |
ollama stop <model> | Stops a running model process. |
ollama show <model> | Displays detailed information about a specific model (parameters, template, etc.). |
ollama serve | Starts the Ollama background service as a local API endpoint. |
ollama create -f Modelfile | Creates a custom model from a Modelfile (this is how you would bake in a SYSTEM prompt). |
⌨️ In-Chat Slash Commands (Type at the >>> Prompt)
Once you are inside a conversation (after running ollama run llama3.2), these commands control the session itself. They are especially useful for managing context.
| Command | What it does |
|---|---|
/help (or /?) | Displays the available in-chat commands. |
/clear | Clears the current conversation context. Use this to make the model “forget” earlier messages, which is useful when switching topics to save memory and avoid confusion. |
/bye (or /exit, /quit) | Exits the chat session and returns you to the PowerShell prompt. |
/model <name> (or /load) | Switches to a different model without leaving the current session. |
/save <name> | Saves your current chat session to a file. |
/show | Shows the current session information or model details. |
""" (Triple Quotes) | While not a slash command, wrapping your text in """ lets you paste multiline input (like code or a long abstract) directly into the chat. |
